Upload folder using huggingface_hub

.gitattributes

@@ -33,3 +33,6 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+wandb/wandb/run-20260607_163220-6s9ihina/run-6s9ihina.wandb filter=lfs diff=lfs merge=lfs -text
+wandb/wandb/run-20260607_185700-zhon9num/run-zhon9num.wandb filter=lfs diff=lfs merge=lfs -text
+wandb/wandb/run-20260608_022840-lftcj17d/run-lftcj17d.wandb filter=lfs diff=lfs merge=lfs -text

checkpoints/best_model_metadata.json

@@ -0,0 +1,27 @@
+{
+  "best_step": 1100,
+  "best_score": 24.940000000000126,
+  "score_name": "mid_train_macro_mean_reward",
+  "stage": "mid_train",
+  "eval_result_path": "/workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/eval/mid_train/step_1100.json",
+  "aggregate": {
+    "stage": "mid_train",
+    "step": 1100,
+    "task": "flappy",
+    "model_alias": "openvla",
+    "fixed_episode_seeds": true,
+    "eval_seed": 42,
+    "total_episodes": 5,
+    "mean_reward": 24.940000000000126,
+    "mean_length": 219.0,
+    "std_reward": 16.381282001113483,
+    "std_length": 133.51404420509476,
+    "task_count": 1,
+    "macro_mean_reward": 24.940000000000126,
+    "macro_mean_length": 219.0,
+    "distributed_eval": false
+  },
+  "per_latency_scores": {
+    "flappy/latency_1": 24.940000000000126
+  }
+}

checkpoints/best_state/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6abc545e6b1abb272b7829b5676f412eebf33cfe0f21ba56368eb619c4610d83
+size 9138230516

checkpoints/best_state/optimizer.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a7d8b83ff4cfdca8a8ca125ea5cb573395aef44106e96ef6376583c9f5a00270
+size 18276874922

checkpoints/best_state/random_states_0.pkl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a71cfec9325ea4c2cb3a6d5bf76c31838dde6cb807961971b9a032cf16ca331d
+size 14821

checkpoints/steps_1500_state/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:76504a6629f8a781c5cfdbeac1795bdc6e221ba0d1931eabaf12d5bf5d56ccc1
+size 9138230516

checkpoints/steps_1500_state/optimizer.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:51b0fcb096b9a9b99a3164ca4b14d9dbd0f82d564c64a330458eee2fdf7a667d
+size 18276874922

checkpoints/steps_1500_state/random_states_0.pkl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:24e6ff8f877c72d822ab74ac8139cb3750c6b68d9b3e4abda7db05b7cc5a0ab8
+size 14821

config.full.yaml

@@ -0,0 +1,187 @@
+framework:
+  name: QwenOFT
+  qwenvl:
+    base_vlm: /workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct
+    attn_implementation: flash_attention_2
+    enable_gradient_checkpointing: true
+  action_model:
+    action_model_type: MLP
+    action_dim: 7
+    action_hidden_dim: 2560
+    future_action_window_size: 0
+    past_action_window_size: 0
+    loss_type: discrete_ce
+    state_dim: 7
+    action_horizon: 1
+    action_env_dim: 2
+datasets:
+  vla_data:
+    dataset_py: lerobot_datasets
+    include_state: true
+    data_root_dir: /workspace/starVLA/data/flappy_fix_latency_1
+    data_mix: flappy_train__bridge
+    eval_data_mix: flappy_train__bridge__val
+    custom_mixtures_path: null
+    action_type: discrete
+    sequential_step_sampling: false
+    per_device_batch_size: 32
+    load_all_data_for_training: true
+    obs_image_size: null
+    video_backend: torchvision_av
+dataset:
+  source_hf: ''
+  converted_name: flappy_train
+  single_source_hf: ''
+  mixed_source_hf: ''
+  single_converted_name: flappy_train
+  mixed_converted_name: flappy_mixed_latency_train
+  single_latency_filter: null
+  mixed_latency_filter: null
+  force_download: false
+  setup_force: false
+  verify_rows: 200
+  max_episodes: null
+  episodes_per_latency: null
+  latency_filter: null
+  debug_subset:
+    enabled: false
+    max_episodes: 5
+    suffix: debug
+base_model:
+  repo_id: Qwen/Qwen3-VL-4B-Instruct
+initialization:
+  checkpoint_local_dir: playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1
+  checkpoint_hf_repo_id: StarVLA/Qwen3VL-OFT-Bridge-RT-1
+  checkpoint_filename: checkpoints/steps_5000_pytorch_model.pt
+trainer:
+  max_train_steps: 1500
+  num_warmup_steps: 100
+  save_interval: 100
+  eval_interval: 100
+  eval_num_batches: 20
+  learning_rate:
+    base: 2.0e-05
+    qwen_vl_interface: 1.0e-05
+    action_model: 0.0001
+  lr_scheduler_type: cosine_with_min_lr
+  scheduler_specific_kwargs:
+    min_lr: 1.0e-06
+  freeze_modules: ''
+  loss_scale:
+    vla: 1.0
+    vlm: 0.1
+  max_grad_norm: 1.0
+  weight_decay: 0.0
+  logging_frequency: 10
+  gradient_clipping: 1.0
+  gradient_accumulation_steps: 16
+  distributed_backend: none
+  is_resume: true
+  pretrained_checkpoint: /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_500_state
+  resume_step: 500
+  reload_modules: null
+  optimizer:
+    name: AdamW
+    betas:
+    - 0.9
+    - 0.95
+    eps: 1.0e-08
+    weight_decay: 1.0e-08
+    fused: false
+  save_format: pt
+workspace_dir: WORKSPACE_DIR
+run_root_dir: /workspace/starVLA/results/Checkpoints
+seed: 42
+wandb_entity: dongqianyu99-zhejiang-university
+wandb_project: starVLA_rl_games
+auth:
+  env_file: null
+  hf_token_env: HF_TOKEN
+  wandb_api_key_env: WANDB_API_KEY
+paths:
+  run_root_dir: results/Checkpoints
+  dataset_local_dir: playground/Datasets/rl_games
+  dataset_cache_dir: null
+  base_model_dir: playground/Pretrained_models/Qwen3-VL-4B-Instruct
+  accelerate_config: starVLA/config/deepseeds/deepspeed_zero2.yaml
+launch:
+  use_accelerate: true
+  gpus: null
+  num_processes: 1
+  dry_run: false
+conda:
+  enabled: true
+  env_name: null
+rl_games:
+  model_alias: openvla
+  env_eval:
+    image_size: 224
+    frameskip: 1
+    seed: 42
+    fixed_episode_seeds: true
+    latency_seed_stride: 0
+    task_seed_stride: 0
+    task_description: ''
+    enabled: true
+    distributed_mode: none
+    vectorized:
+      enabled: false
+      batch_size: 1
+    latency:
+      prompt_map_path: /workspace/starVLA/data/flappy_fix_latency_1/flappy_train__bridge/latency_prompt_map.json
+      mode: single
+      values:
+      - 0
+    mid_train:
+      enabled: true
+      interval_steps: 100
+      latencies:
+      - 1
+      num_episodes: 5
+      max_steps_per_episode: 3600
+    post_train:
+      enabled: true
+      latencies:
+      - 0
+      - 1
+      - 2
+      - 3
+      - 4
+      num_episodes: 5
+      max_steps_per_episode: 3600
+  task: flappy
+  initialization_mode: bridge
+  action_carrier: bridge
+model: openvla
+env: flappy
+init: bridge
+bridge_base_model:
+  repo_id:
+    openvla: Qwen/Qwen3-VL-4B-Instruct
+    pi0: StarVLA/Qwen2.5-VL-3B-Instruct-Action
+    pi05: Qwen/Qwen3-VL-4B-Instruct
+    gr00t: Qwen/Qwen3-VL-4B-Instruct
+  local_dir:
+    openvla: playground/Pretrained_models/Qwen3-VL-4B-Instruct
+    pi0: playground/Pretrained_models/Qwen2.5-VL-3B-Instruct-Action
+    pi05: playground/Pretrained_models/Qwen3-VL-4B-Instruct
+    gr00t: playground/Pretrained_models/Qwen3-VL-4B-Instruct
+mode: single
+checkpoint:
+  load: auto
+  hf_repo_id: null
+  save_best_model: true
+  save_pt_file: false
+  local:
+    keep_last_n: 1
+  sync:
+    enabled: false
+    repo_id: null
+    keep_last_n: 0
+    sync_every_n_checkpoints: 1
+    resume_policy: local_latest
+run_id: openvla_flappy_fix_latency_1
+output_dir: /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1
+config_yaml: null
+is_debug: false
+version_id: '0.21'

config.yaml

@@ -0,0 +1,87 @@
+checkpoint:
+  local:
+    keep_last_n: 1
+  save_best_model: true
+  save_pt_file: false
+  sync:
+    enabled: false
+    keep_last_n: 0
+    repo_id: null
+datasets:
+  vla_data:
+    data_mix: flappy_train__bridge
+    dataset_py: lerobot_datasets
+    eval_data_mix: flappy_train__bridge__val
+    include_state: true
+    obs_image_size: null
+    per_device_batch_size: 32
+framework:
+  action_model:
+    action_dim: 7
+    action_env_dim: 2
+    action_hidden_dim: 2560
+    action_horizon: 1
+    action_model_type: MLP
+    loss_type: discrete_ce
+    state_dim: 7
+  name: QwenOFT
+  qwenvl:
+    attn_implementation: flash_attention_2
+    base_vlm: /workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct
+    enable_gradient_checkpointing: true
+output_dir: /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1
+rl_games:
+  env_eval:
+    distributed_mode: none
+    enabled: true
+    fixed_episode_seeds: true
+    frameskip: 1
+    image_size: 224
+    latency:
+      prompt_map_path: /workspace/starVLA/data/flappy_fix_latency_1/flappy_train__bridge/latency_prompt_map.json
+    latency_seed_stride: 0
+    mid_train:
+      enabled: true
+      interval_steps: 100
+      latencies:
+      - 1
+      max_steps_per_episode: 3600
+      num_episodes: 5
+    seed: 42
+    task_description: ''
+    task_seed_stride: 0
+    vectorized:
+      enabled: false
+  model_alias: openvla
+  task: flappy
+run_id: openvla_flappy_fix_latency_1
+run_root_dir: /workspace/starVLA/results/Checkpoints
+seed: 42
+trainer:
+  eval_interval: 100
+  eval_num_batches: 20
+  freeze_modules: ''
+  gradient_clipping: 1.0
+  is_resume: true
+  learning_rate:
+    action_model: 0.0001
+    base: 2.0e-05
+    qwen_vl_interface: 1.0e-05
+  logging_frequency: 10
+  lr_scheduler_type: cosine_with_min_lr
+  max_train_steps: 1500
+  num_warmup_steps: 100
+  optimizer:
+    betas:
+    - 0.9
+    - 0.95
+    eps: 1.0e-08
+    fused: false
+    weight_decay: 1.0e-08
+  pretrained_checkpoint: /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_500_state
+  resume_step: 500
+  save_interval: 100
+  scheduler_specific_kwargs:
+    min_lr: 1.0e-06
+wandb_entity: dongqianyu99-zhejiang-university
+wandb_project: starVLA_rl_games

dataset_statistics.json

@@ -0,0 +1,127 @@
+{
+  "new_embodiment": {
+    "action": {
+      "mean": [
+        0.8599907159805298,
+        0.14000925421714783,
+        0.0,
+        0.0,
+        0.0,
+        0.0,
+        0.0
+      ],
+      "std": [
+        0.34691014885902405,
+        0.34691014885902405,
+        0.0,
+        0.0,
+        0.0,
+        0.0,
+        0.0
+      ],
+      "max": [
+        1.0,
+        1.0,
+        0.0,
+        0.0,
+        0.0,
+        0.0,
+        0.0
+      ],
+      "min": [
+        0.0,
+        0.0,
+        0.0,
+        0.0,
+        0.0,
+        0.0,
+        0.0
+      ],
+      "q01": [
+        0.0,
+        0.0,
+        0.0,
+        0.0,
+        0.0,
+        0.0,
+        0.0
+      ],
+      "q99": [
+        1.0,
+        1.0,
+        0.0,
+        0.0,
+        0.0,
+        0.0,
+        0.0
+      ],
+      "mask": [
+        true,
+        true,
+        true,
+        true,
+        true,
+        true,
+        true
+      ]
+    },
+    "state": {
+      "mean": [
+        0.0,
+        0.0,
+        0.0,
+        0.0,
+        0.0,
+        0.0,
+        0.0
+      ],
+      "std": [
+        0.0,
+        0.0,
+        0.0,
+        0.0,
+        0.0,
+        0.0,
+        0.0
+      ],
+      "max": [
+        0.0,
+        0.0,
+        0.0,
+        0.0,
+        0.0,
+        0.0,
+        0.0
+      ],
+      "min": [
+        0.0,
+        0.0,
+        0.0,
+        0.0,
+        0.0,
+        0.0,
+        0.0
+      ],
+      "q01": [
+        0.0,
+        0.0,
+        0.0,
+        0.0,
+        0.0,
+        0.0,
+        0.0
+      ],
+      "q99": [
+        0.0,
+        0.0,
+        0.0,
+        0.0,
+        0.0,
+        0.0,
+        0.0
+      ]
+    },
+    "num_transitions": 108000,
+    "num_trajectories": 30
+  }
+}

dataset_statistics_eval.json

@@ -0,0 +1,127 @@
+{
+  "new_embodiment": {
+    "action": {
+      "mean": [
+        0.8598055839538574,
+        0.14019444584846497,
+        0.0,
+        0.0,
+        0.0,
+        0.0,
+        0.0
+      ],
+      "std": [
+        0.3472478985786438,
+        0.3472478985786438,
+        0.0,
+        0.0,
+        0.0,
+        0.0,
+        0.0
+      ],
+      "max": [
+        1.0,
+        1.0,
+        0.0,
+        0.0,
+        0.0,
+        0.0,
+        0.0
+      ],
+      "min": [
+        0.0,
+        0.0,
+        0.0,
+        0.0,
+        0.0,
+        0.0,
+        0.0
+      ],
+      "q01": [
+        0.0,
+        0.0,
+        0.0,
+        0.0,
+        0.0,
+        0.0,
+        0.0
+      ],
+      "q99": [
+        1.0,
+        1.0,
+        0.0,
+        0.0,
+        0.0,
+        0.0,
+        0.0
+      ],
+      "mask": [
+        true,
+        true,
+        true,
+        true,
+        true,
+        true,
+        true
+      ]
+    },
+    "state": {
+      "mean": [
+        0.0,
+        0.0,
+        0.0,
+        0.0,
+        0.0,
+        0.0,
+        0.0
+      ],
+      "std": [
+        0.0,
+        0.0,
+        0.0,
+        0.0,
+        0.0,
+        0.0,
+        0.0
+      ],
+      "max": [
+        0.0,
+        0.0,
+        0.0,
+        0.0,
+        0.0,
+        0.0,
+        0.0
+      ],
+      "min": [
+        0.0,
+        0.0,
+        0.0,
+        0.0,
+        0.0,
+        0.0,
+        0.0
+      ],
+      "q01": [
+        0.0,
+        0.0,
+        0.0,
+        0.0,
+        0.0,
+        0.0,
+        0.0
+      ],
+      "q99": [
+        0.0,
+        0.0,
+        0.0,
+        0.0,
+        0.0,
+        0.0,
+        0.0
+      ]
+    },
+    "num_transitions": 36000,
+    "num_trajectories": 10
+  }
+}

hydra/.hydra/config.yaml

@@ -0,0 +1,185 @@
+framework:
+  qwenvl:
+    base_vlm: /workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct
+    attn_implementation: flash_attention_2
+    enable_gradient_checkpointing: true
+  action_model:
+    state_dim: 7
+    loss_type: discrete_ce
+    action_horizon: 1
+    future_action_window_size: 0
+    past_action_window_size: 0
+    action_dim: 7
+    action_env_dim: 2
+  name: QwenOFT
+datasets:
+  vla_data:
+    dataset_py: lerobot_datasets
+    include_state: true
+    data_root_dir: /workspace/starVLA/data/flappy_fix_latency_1
+    data_mix: flappy_train__bridge
+    eval_data_mix: flappy_train__bridge__val
+    custom_mixtures_path: null
+    action_type: discrete
+    sequential_step_sampling: false
+    per_device_batch_size: 32
+    load_all_data_for_training: true
+    obs_image_size: null
+    video_backend: torchvision_av
+dataset:
+  source_hf: ${dataset.single_source_hf}
+  converted_name: ${dataset.single_converted_name}
+  single_source_hf: ''
+  mixed_source_hf: ''
+  single_converted_name: flappy_train
+  mixed_converted_name: flappy_mixed_latency_train
+  single_latency_filter: null
+  mixed_latency_filter: null
+  force_download: false
+  setup_force: false
+  verify_rows: 200
+  max_episodes: null
+  episodes_per_latency: null
+  latency_filter: ${dataset.single_latency_filter}
+  debug_subset:
+    enabled: false
+    max_episodes: 5
+    suffix: debug
+base_model:
+  repo_id: ${bridge_base_model.repo_id.${model}}
+initialization:
+  checkpoint_local_dir: playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1
+  checkpoint_hf_repo_id: StarVLA/Qwen3VL-OFT-Bridge-RT-1
+  checkpoint_filename: checkpoints/steps_5000_pytorch_model.pt
+trainer:
+  max_train_steps: 1500
+  num_warmup_steps: 100
+  save_interval: 100
+  eval_interval: 100
+  eval_num_batches: 20
+  learning_rate:
+    base: 2.0e-05
+    qwen_vl_interface: 1.0e-05
+    action_model: 0.0001
+  lr_scheduler_type: cosine_with_min_lr
+  scheduler_specific_kwargs:
+    min_lr: 1.0e-06
+  freeze_modules: ''
+  loss_scale:
+    vla: 1.0
+    vlm: 0.1
+  max_grad_norm: 1.0
+  weight_decay: 0.0
+  logging_frequency: 10
+  gradient_clipping: 1.0
+  gradient_accumulation_steps: 16
+  distributed_backend: none
+  is_resume: true
+  pretrained_checkpoint: /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_500_state
+  resume_step: 500
+  reload_modules: null
+  optimizer:
+    name: AdamW
+    betas:
+    - 0.9
+    - 0.95
+    eps: 1.0e-08
+    weight_decay: 1.0e-08
+    fused: false
+  save_format: pt
+workspace_dir: WORKSPACE_DIR
+run_root_dir: /workspace/starVLA/results/Checkpoints
+seed: 42
+wandb_entity: dongqianyu99-zhejiang-university
+wandb_project: starVLA_rl_games
+auth:
+  env_file: null
+  hf_token_env: HF_TOKEN
+  wandb_api_key_env: WANDB_API_KEY
+paths:
+  run_root_dir: results/Checkpoints
+  dataset_local_dir: playground/Datasets/rl_games
+  dataset_cache_dir: null
+  base_model_dir: ${bridge_base_model.local_dir.${model}}
+  accelerate_config: starVLA/config/deepseeds/deepspeed_zero2.yaml
+launch:
+  use_accelerate: true
+  gpus: null
+  num_processes: 1
+  dry_run: false
+conda:
+  enabled: true
+  env_name: null
+rl_games:
+  model_alias: openvla
+  env_eval:
+    image_size: 224
+    frameskip: 1
+    seed: 42
+    fixed_episode_seeds: true
+    latency_seed_stride: 0
+    task_seed_stride: 0
+    task_description: ''
+    enabled: true
+    distributed_mode: none
+    vectorized:
+      enabled: false
+      batch_size: 1
+    latency:
+      prompt_map_path: /workspace/starVLA/data/flappy_fix_latency_1/flappy_train__bridge/latency_prompt_map.json
+      mode: single
+      values:
+      - 0
+    mid_train:
+      enabled: true
+      interval_steps: 100
+      latencies:
+      - 1
+      num_episodes: 5
+      max_steps_per_episode: 3600
+    post_train:
+      enabled: true
+      latencies:
+      - 0
+      - 1
+      - 2
+      - 3
+      - 4
+      num_episodes: 5
+      max_steps_per_episode: 3600
+  task: flappy
+  initialization_mode: bridge
+  action_carrier: bridge
+model: openvla
+env: flappy
+init: bridge
+bridge_base_model:
+  repo_id:
+    openvla: Qwen/Qwen3-VL-4B-Instruct
+    pi0: StarVLA/Qwen2.5-VL-3B-Instruct-Action
+    pi05: Qwen/Qwen3-VL-4B-Instruct
+    gr00t: Qwen/Qwen3-VL-4B-Instruct
+  local_dir:
+    openvla: playground/Pretrained_models/Qwen3-VL-4B-Instruct
+    pi0: playground/Pretrained_models/Qwen2.5-VL-3B-Instruct-Action
+    pi05: playground/Pretrained_models/Qwen3-VL-4B-Instruct
+    gr00t: playground/Pretrained_models/Qwen3-VL-4B-Instruct
+mode: single
+checkpoint:
+  load: auto
+  hf_repo_id: null
+  save_best_model: true
+  save_pt_file: false
+  local:
+    keep_last_n: 1
+  sync:
+    enabled: false
+    repo_id: null
+    keep_last_n: 0
+    sync_every_n_checkpoints: 1
+    resume_policy: local_latest
+run_id: openvla_flappy_fix_latency_1
+output_dir: null
+config_yaml: null
+is_debug: false
+version_id: '0.21'

hydra/.hydra/hydra.yaml

@@ -0,0 +1,246 @@
+hydra:
+  run:
+    dir: ${run_root_dir}/${run_id}/hydra
+  sweep:
+    dir: multirun/${now:%Y-%m-%d}/${now:%H-%M-%S}
+    subdir: ${hydra.job.num}
+  launcher:
+    _target_: hydra._internal.core_plugins.basic_launcher.BasicLauncher
+  sweeper:
+    _target_: hydra._internal.core_plugins.basic_sweeper.BasicSweeper
+    max_batch_size: null
+    params: null
+  help:
+    app_name: ${hydra.job.name}
+    header: '${hydra.help.app_name} is powered by Hydra.
+
+      '
+    footer: 'Powered by Hydra (https://hydra.cc)
+
+      Use --hydra-help to view Hydra specific help
+
+      '
+    template: '${hydra.help.header}
+
+      == Configuration groups ==
+
+      Compose your configuration from those groups (group=option)
+
+
+      $APP_CONFIG_GROUPS
+
+
+      == Config ==
+
+      Override anything in the config (foo.bar=value)
+
+
+      $CONFIG
+
+
+      ${hydra.help.footer}
+
+      '
+  hydra_help:
+    template: 'Hydra (${hydra.runtime.version})
+
+      See https://hydra.cc for more info.
+
+
+      == Flags ==
+
+      $FLAGS_HELP
+
+
+      == Configuration groups ==
+
+      Compose your configuration from those groups (For example, append hydra/job_logging=disabled
+      to command line)
+
+
+      $HYDRA_CONFIG_GROUPS
+
+
+      Use ''--cfg hydra'' to Show the Hydra config.
+
+      '
+    hydra_help: ???
+  hydra_logging:
+    version: 1
+    formatters:
+      simple:
+        format: '[%(asctime)s][HYDRA] %(message)s'
+    handlers:
+      console:
+        class: logging.StreamHandler
+        formatter: simple
+        stream: ext://sys.stdout
+    root:
+      level: INFO
+      handlers:
+      - console
+    loggers:
+      logging_example:
+        level: DEBUG
+    disable_existing_loggers: false
+  job_logging:
+    version: 1
+    formatters:
+      simple:
+        format: '[%(asctime)s][%(name)s][%(levelname)s] - %(message)s'
+    handlers:
+      console:
+        class: logging.StreamHandler
+        formatter: simple
+        stream: ext://sys.stdout
+      file:
+        class: logging.FileHandler
+        formatter: simple
+        filename: ${hydra.runtime.output_dir}/${hydra.job.name}.log
+    root:
+      level: INFO
+      handlers:
+      - console
+      - file
+    disable_existing_loggers: false
+  env: {}
+  mode: RUN
+  searchpath: []
+  callbacks: {}
+  output_subdir: .hydra
+  overrides:
+    hydra:
+    - hydra.mode=RUN
+    task:
+    - model=openvla
+    - env=flappy
+    - init=bridge
+    - mode=single
+    - run_id=openvla_flappy_fix_latency_1
+    - run_root_dir=/workspace/starVLA/results/Checkpoints
+    - seed=42
+    - wandb_entity=dongqianyu99-zhejiang-university
+    - wandb_project=starVLA_rl_games
+    - rl_games.env_eval.enabled=true
+    - checkpoint.sync.enabled=false
+    - checkpoint.sync.keep_last_n=0
+    - checkpoint.local.keep_last_n=1
+    - checkpoint.save_best_model=true
+    - checkpoint.save_pt_file=false
+    - trainer.is_resume=true
+    - trainer.pretrained_checkpoint=/workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_500_state
+    - trainer.resume_step=500
+    - trainer.max_train_steps=1500
+    - trainer.num_warmup_steps=100
+    - trainer.save_interval=100
+    - trainer.eval_interval=100
+    - trainer.eval_num_batches=20
+    - trainer.logging_frequency=10
+    - trainer.gradient_accumulation_steps=16
+    - trainer.distributed_backend=none
+    - trainer.learning_rate.base=2e-05
+    - trainer.learning_rate.qwen_vl_interface=1e-05
+    - trainer.learning_rate.action_model=0.0001
+    - trainer.lr_scheduler_type=cosine_with_min_lr
+    - trainer.scheduler_specific_kwargs.min_lr=1e-06
+    - trainer.loss_scale.vla=1.0
+    - trainer.loss_scale.vlm=0.1
+    - trainer.max_grad_norm=1.0
+    - trainer.weight_decay=0.0
+    - trainer.gradient_clipping=1.0
+    - trainer.optimizer.name=AdamW
+    - trainer.optimizer.betas=[0.9, 0.95]
+    - trainer.optimizer.eps=1e-08
+    - trainer.optimizer.weight_decay=1e-08
+    - trainer.optimizer.fused=false
+    - trainer.save_format=pt
+    - framework.name=QwenOFT
+    - framework.qwenvl.attn_implementation=flash_attention_2
+    - framework.qwenvl.enable_gradient_checkpointing=true
+    - framework.action_model.action_dim=7
+    - framework.action_model.action_env_dim=2
+    - framework.action_model.state_dim=7
+    - framework.action_model.loss_type=discrete_ce
+    - framework.action_model.action_horizon=1
+    - framework.action_model.future_action_window_size=0
+    - framework.action_model.past_action_window_size=0
+    - datasets.vla_data.include_state=true
+    - datasets.vla_data.action_type=discrete
+    - datasets.vla_data.sequential_step_sampling=false
+    - datasets.vla_data.per_device_batch_size=32
+    - datasets.vla_data.load_all_data_for_training=true
+    - datasets.vla_data.video_backend=torchvision_av
+    - datasets.vla_data.data_root_dir=/workspace/starVLA/data/flappy_fix_latency_1
+    - datasets.vla_data.data_mix=flappy_train__bridge
+    - datasets.vla_data.eval_data_mix=flappy_train__bridge__val
+    - framework.qwenvl.base_vlm=/workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct
+    - rl_games.task=flappy
+    - rl_games.model_alias=openvla
+    - rl_games.initialization_mode=bridge
+    - rl_games.action_carrier=bridge
+    - rl_games.env_eval.distributed_mode=none
+    - rl_games.env_eval.latency.mode=single
+    - rl_games.env_eval.frameskip=1
+    - rl_games.env_eval.image_size=224
+    - rl_games.env_eval.seed=42
+    - rl_games.env_eval.fixed_episode_seeds=true
+    - rl_games.env_eval.latency_seed_stride=0
+    - rl_games.env_eval.task_seed_stride=0
+    - rl_games.env_eval.latency.values=[0]
+    - rl_games.env_eval.latency.prompt_map_path=/workspace/starVLA/data/flappy_fix_latency_1/flappy_train__bridge/latency_prompt_map.json
+    - rl_games.env_eval.mid_train.enabled=true
+    - rl_games.env_eval.mid_train.interval_steps=100
+    - rl_games.env_eval.mid_train.num_episodes=5
+    - rl_games.env_eval.mid_train.max_steps_per_episode=3600
+    - rl_games.env_eval.mid_train.latencies=[1]
+    - rl_games.env_eval.post_train.enabled=true
+    - rl_games.env_eval.post_train.num_episodes=5
+    - rl_games.env_eval.post_train.max_steps_per_episode=3600
+    - rl_games.env_eval.post_train.latencies=[0,1,2,3,4]
+  job:
+    name: train_starvla_hydra
+    chdir: false
+    override_dirname: checkpoint.local.keep_last_n=1,checkpoint.save_best_model=true,checkpoint.save_pt_file=false,checkpoint.sync.enabled=false,checkpoint.sync.keep_last_n=0,datasets.vla_data.action_type=discrete,datasets.vla_data.data_mix=flappy_train__bridge,datasets.vla_data.data_root_dir=/workspace/starVLA/data/flappy_fix_latency_1,datasets.vla_data.eval_data_mix=flappy_train__bridge__val,datasets.vla_data.include_state=true,datasets.vla_data.load_all_data_for_training=true,datasets.vla_data.per_device_batch_size=32,datasets.vla_data.sequential_step_sampling=false,datasets.vla_data.video_backend=torchvision_av,env=flappy,framework.action_model.action_dim=7,framework.action_model.action_env_dim=2,framework.action_model.action_horizon=1,framework.action_model.future_action_window_size=0,framework.action_model.loss_type=discrete_ce,framework.action_model.past_action_window_size=0,framework.action_model.state_dim=7,framework.name=QwenOFT,framework.qwenvl.attn_implementation=flash_attention_2,framework.qwenvl.base_vlm=/workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct,framework.qwenvl.enable_gradient_checkpointing=true,init=bridge,mode=single,model=openvla,rl_games.action_carrier=bridge,rl_games.env_eval.distributed_mode=none,rl_games.env_eval.enabled=true,rl_games.env_eval.fixed_episode_seeds=true,rl_games.env_eval.frameskip=1,rl_games.env_eval.image_size=224,rl_games.env_eval.latency.mode=single,rl_games.env_eval.latency.prompt_map_path=/workspace/starVLA/data/flappy_fix_latency_1/flappy_train__bridge/latency_prompt_map.json,rl_games.env_eval.latency.values=[0],rl_games.env_eval.latency_seed_stride=0,rl_games.env_eval.mid_train.enabled=true,rl_games.env_eval.mid_train.interval_steps=100,rl_games.env_eval.mid_train.latencies=[1],rl_games.env_eval.mid_train.max_steps_per_episode=3600,rl_games.env_eval.mid_train.num_episodes=5,rl_games.env_eval.post_train.enabled=true,rl_games.env_eval.post_train.latencies=[0,1,2,3,4],rl_games.env_eval.post_train.max_steps_per_episode=3600,rl_games.env_eval.post_train.num_episodes=5,rl_games.env_eval.seed=42,rl_games.env_eval.task_seed_stride=0,rl_games.initialization_mode=bridge,rl_games.model_alias=openvla,rl_games.task=flappy,run_id=openvla_flappy_fix_latency_1,run_root_dir=/workspace/starVLA/results/Checkpoints,seed=42,trainer.distributed_backend=none,trainer.eval_interval=100,trainer.eval_num_batches=20,trainer.gradient_accumulation_steps=16,trainer.gradient_clipping=1.0,trainer.is_resume=true,trainer.learning_rate.action_model=0.0001,trainer.learning_rate.base=2e-05,trainer.learning_rate.qwen_vl_interface=1e-05,trainer.logging_frequency=10,trainer.loss_scale.vla=1.0,trainer.loss_scale.vlm=0.1,trainer.lr_scheduler_type=cosine_with_min_lr,trainer.max_grad_norm=1.0,trainer.max_train_steps=1500,trainer.num_warmup_steps=100,trainer.optimizer.betas=[0.9,
+      0.95],trainer.optimizer.eps=1e-08,trainer.optimizer.fused=false,trainer.optimizer.name=AdamW,trainer.optimizer.weight_decay=1e-08,trainer.pretrained_checkpoint=/workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_500_state,trainer.resume_step=500,trainer.save_format=pt,trainer.save_interval=100,trainer.scheduler_specific_kwargs.min_lr=1e-06,trainer.weight_decay=0.0,wandb_entity=dongqianyu99-zhejiang-university,wandb_project=starVLA_rl_games
+    id: ???
+    num: ???
+    config_name: train
+    env_set: {}
+    env_copy: []
+    config:
+      override_dirname:
+        kv_sep: '='
+        item_sep: ','
+        exclude_keys: []
+  runtime:
+    version: 1.3.2
+    version_base: '1.1'
+    cwd: /workspace/starVLA
+    config_sources:
+    - path: hydra.conf
+      schema: pkg
+      provider: hydra
+    - path: /workspace/starVLA/examples/rl_games/config
+      schema: file
+      provider: main
+    - path: ''
+      schema: structured
+      provider: schema
+    output_dir: /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/hydra
+    choices:
+      cross_task_setup: null
+      checkpoint: default
+      mode: single
+      init: bridge
+      env: flappy
+      model: openvla
+      hydra/env: default
+      hydra/callbacks: null
+      hydra/job_logging: default
+      hydra/hydra_logging: default
+      hydra/hydra_help: default
+      hydra/help: default
+      hydra/sweeper: basic
+      hydra/launcher: basic
+      hydra/output: default
+  verbose: false

hydra/.hydra/overrides.yaml

@@ -0,0 +1,85 @@
+- model=openvla
+- env=flappy
+- init=bridge
+- mode=single
+- run_id=openvla_flappy_fix_latency_1
+- run_root_dir=/workspace/starVLA/results/Checkpoints
+- seed=42
+- wandb_entity=dongqianyu99-zhejiang-university
+- wandb_project=starVLA_rl_games
+- rl_games.env_eval.enabled=true
+- checkpoint.sync.enabled=false
+- checkpoint.sync.keep_last_n=0
+- checkpoint.local.keep_last_n=1
+- checkpoint.save_best_model=true
+- checkpoint.save_pt_file=false
+- trainer.is_resume=true
+- trainer.pretrained_checkpoint=/workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_500_state
+- trainer.resume_step=500
+- trainer.max_train_steps=1500
+- trainer.num_warmup_steps=100
+- trainer.save_interval=100
+- trainer.eval_interval=100
+- trainer.eval_num_batches=20
+- trainer.logging_frequency=10
+- trainer.gradient_accumulation_steps=16
+- trainer.distributed_backend=none
+- trainer.learning_rate.base=2e-05
+- trainer.learning_rate.qwen_vl_interface=1e-05
+- trainer.learning_rate.action_model=0.0001
+- trainer.lr_scheduler_type=cosine_with_min_lr
+- trainer.scheduler_specific_kwargs.min_lr=1e-06
+- trainer.loss_scale.vla=1.0
+- trainer.loss_scale.vlm=0.1
+- trainer.max_grad_norm=1.0
+- trainer.weight_decay=0.0
+- trainer.gradient_clipping=1.0
+- trainer.optimizer.name=AdamW
+- trainer.optimizer.betas=[0.9, 0.95]
+- trainer.optimizer.eps=1e-08
+- trainer.optimizer.weight_decay=1e-08
+- trainer.optimizer.fused=false
+- trainer.save_format=pt
+- framework.name=QwenOFT
+- framework.qwenvl.attn_implementation=flash_attention_2
+- framework.qwenvl.enable_gradient_checkpointing=true
+- framework.action_model.action_dim=7
+- framework.action_model.action_env_dim=2
+- framework.action_model.state_dim=7
+- framework.action_model.loss_type=discrete_ce
+- framework.action_model.action_horizon=1
+- framework.action_model.future_action_window_size=0
+- framework.action_model.past_action_window_size=0
+- datasets.vla_data.include_state=true
+- datasets.vla_data.action_type=discrete
+- datasets.vla_data.sequential_step_sampling=false
+- datasets.vla_data.per_device_batch_size=32
+- datasets.vla_data.load_all_data_for_training=true
+- datasets.vla_data.video_backend=torchvision_av
+- datasets.vla_data.data_root_dir=/workspace/starVLA/data/flappy_fix_latency_1
+- datasets.vla_data.data_mix=flappy_train__bridge
+- datasets.vla_data.eval_data_mix=flappy_train__bridge__val
+- framework.qwenvl.base_vlm=/workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct
+- rl_games.task=flappy
+- rl_games.model_alias=openvla
+- rl_games.initialization_mode=bridge
+- rl_games.action_carrier=bridge
+- rl_games.env_eval.distributed_mode=none
+- rl_games.env_eval.latency.mode=single
+- rl_games.env_eval.frameskip=1
+- rl_games.env_eval.image_size=224
+- rl_games.env_eval.seed=42
+- rl_games.env_eval.fixed_episode_seeds=true
+- rl_games.env_eval.latency_seed_stride=0
+- rl_games.env_eval.task_seed_stride=0
+- rl_games.env_eval.latency.values=[0]
+- rl_games.env_eval.latency.prompt_map_path=/workspace/starVLA/data/flappy_fix_latency_1/flappy_train__bridge/latency_prompt_map.json
+- rl_games.env_eval.mid_train.enabled=true
+- rl_games.env_eval.mid_train.interval_steps=100
+- rl_games.env_eval.mid_train.num_episodes=5
+- rl_games.env_eval.mid_train.max_steps_per_episode=3600
+- rl_games.env_eval.mid_train.latencies=[1]
+- rl_games.env_eval.post_train.enabled=true
+- rl_games.env_eval.post_train.num_episodes=5
+- rl_games.env_eval.post_train.max_steps_per_episode=3600
+- rl_games.env_eval.post_train.latencies=[0,1,2,3,4]

summary.jsonl

@@ -0,0 +1,15 @@
+{"steps": 100}
+{"steps": 200}
+{"steps": 300}
+{"steps": 400}
+{"steps": 500}
+{"steps": 600}
+{"steps": 700}
+{"steps": 800}
+{"steps": 900}
+{"steps": 1000}
+{"steps": 1100}
+{"steps": 1200}
+{"steps": 1300}
+{"steps": 1400}
+{"steps": 1500}

wandb/wandb/debug.log

@@ -0,0 +1,24 @@
+2026-06-08 02:28:40,756 INFO    MainThread:2222 [wandb_setup.py:_flush():81] Current SDK version is 0.27.0
+2026-06-08 02:28:40,756 INFO    MainThread:2222 [wandb_setup.py:_flush():81] Configure stats pid to 2222
+2026-06-08 02:28:40,756 INFO    MainThread:2222 [wandb_setup.py:_flush():81] Loading settings from environment variables
+2026-06-08 02:28:40,757 INFO    MainThread:2222 [wandb_init.py:setup_run_log_directory():723] Logging user logs to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/wandb/wandb/run-20260608_022840-lftcj17d/logs/debug.log
+2026-06-08 02:28:40,757 INFO    MainThread:2222 [wandb_init.py:setup_run_log_directory():724] Logging internal logs to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/wandb/wandb/run-20260608_022840-lftcj17d/logs/debug-internal.log
+2026-06-08 02:28:40,757 INFO    MainThread:2222 [wandb_init.py:init():851] calling init triggers
+2026-06-08 02:28:40,757 INFO    MainThread:2222 [wandb_init.py:init():856] wandb.init called with sweep_config: {}
+config: {'_wandb': {}}
+2026-06-08 02:28:40,757 INFO    MainThread:2222 [wandb_init.py:init():899] starting backend
+2026-06-08 02:28:41,015 INFO    MainThread:2222 [wandb_init.py:init():914] sending inform_init request
+2026-06-08 02:28:41,321 INFO    MainThread:2222 [wandb_init.py:init():919] backend started and connected
+2026-06-08 02:28:41,326 INFO    MainThread:2222 [wandb_init.py:init():989] updated telemetry
+2026-06-08 02:28:41,333 INFO    MainThread:2222 [wandb_init.py:init():1012] communicating run to backend with 90.0 second timeout
+2026-06-08 02:28:41,627 INFO    MainThread:2222 [wandb_init.py:init():1057] starting run threads in backend
+2026-06-08 02:28:41,709 INFO    MainThread:2222 [wandb_run.py:_console_start():2509] atexit reg
+2026-06-08 02:28:41,709 INFO    MainThread:2222 [wandb_run.py:_redirect():2359] redirect: wrap_raw
+2026-06-08 02:28:41,709 INFO    MainThread:2222 [wandb_run.py:_redirect():2428] Wrapping output streams.
+2026-06-08 02:28:41,709 INFO    MainThread:2222 [wandb_run.py:_redirect():2451] Redirects installed.
+2026-06-08 02:28:41,712 INFO    MainThread:2222 [wandb_init.py:init():1095] run started, returning control to user process
+2026-06-08 08:13:05,689 INFO    MainThread:2222 [wandb_run.py:_finish():2271] finishing run dongqianyu99-zhejiang-university/starVLA_rl_games/lftcj17d
+2026-06-08 08:13:05,690 INFO    MainThread:2222 [wandb_run.py:_atexit_cleanup():2476] got exitcode: 0
+2026-06-08 08:13:05,690 INFO    MainThread:2222 [wandb_run.py:_restore():2458] restore
+2026-06-08 08:13:05,690 INFO    MainThread:2222 [wandb_run.py:_restore():2464] restore done
+2026-06-08 08:13:07,616 INFO    MainThread:2222 [wandb_run.py:_footer_sync_info():3877] logging synced files

wandb/wandb/run-20260607_163058-ef6e3dz3/files/config.yaml

@@ -0,0 +1,155 @@
+_wandb:
+    value:
+        cli_version: 0.27.0
+        e:
+            rhxgvu8eyjiqv9yk50aggi440ai2lafp:
+                args:
+                    - --config-name
+                    - train
+                    - model=openvla
+                    - env=flappy
+                    - init=bridge
+                    - mode=single
+                    - run_id=openvla_flappy_fix_latency_1
+                    - run_root_dir=/workspace/starVLA/results/Checkpoints
+                    - seed=42
+                    - wandb_entity=dongqianyu99-zhejiang-university
+                    - wandb_project=starVLA_rl_games
+                    - rl_games.env_eval.enabled=true
+                    - checkpoint.sync.enabled=false
+                    - checkpoint.sync.keep_last_n=0
+                    - checkpoint.local.keep_last_n=1
+                    - checkpoint.save_best_model=true
+                    - checkpoint.save_pt_file=false
+                    - trainer.is_resume=false
+                    - trainer.pretrained_checkpoint=/workspace/starVLA/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt
+                    - trainer.resume_step=0
+                    - trainer.max_train_steps=1500
+                    - trainer.num_warmup_steps=100
+                    - trainer.save_interval=100
+                    - trainer.eval_interval=100
+                    - trainer.eval_num_batches=20
+                    - trainer.logging_frequency=10
+                    - trainer.gradient_accumulation_steps=16
+                    - trainer.distributed_backend=none
+                    - trainer.learning_rate.base=2e-05
+                    - trainer.learning_rate.qwen_vl_interface=1e-05
+                    - trainer.learning_rate.action_model=0.0001
+                    - trainer.lr_scheduler_type=cosine_with_min_lr
+                    - trainer.scheduler_specific_kwargs.min_lr=1e-06
+                    - trainer.loss_scale.vla=1.0
+                    - trainer.loss_scale.vlm=0.1
+                    - trainer.max_grad_norm=1.0
+                    - trainer.weight_decay=0.0
+                    - trainer.gradient_clipping=1.0
+                    - trainer.optimizer.name=AdamW
+                    - trainer.optimizer.betas=[0.9, 0.95]
+                    - trainer.optimizer.eps=1e-08
+                    - trainer.optimizer.weight_decay=1e-08
+                    - trainer.optimizer.fused=false
+                    - trainer.save_format=pt
+                    - framework.name=QwenOFT
+                    - framework.qwenvl.attn_implementation=flash_attention_2
+                    - framework.qwenvl.enable_gradient_checkpointing=true
+                    - framework.action_model.action_dim=7
+                    - framework.action_model.action_env_dim=2
+                    - framework.action_model.state_dim=7
+                    - framework.action_model.loss_type=discrete_ce
+                    - framework.action_model.action_horizon=1
+                    - framework.action_model.future_action_window_size=0
+                    - framework.action_model.past_action_window_size=0
+                    - datasets.vla_data.include_state=true
+                    - datasets.vla_data.action_type=discrete
+                    - datasets.vla_data.sequential_step_sampling=false
+                    - datasets.vla_data.per_device_batch_size=32
+                    - datasets.vla_data.load_all_data_for_training=true
+                    - datasets.vla_data.video_backend=torchvision_av
+                    - datasets.vla_data.data_root_dir=/workspace/starVLA/data/flappy_fix_latency_1
+                    - datasets.vla_data.data_mix=flappy_train__bridge
+                    - datasets.vla_data.eval_data_mix=flappy_train__bridge__val
+                    - framework.qwenvl.base_vlm=/workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct
+                    - rl_games.task=flappy
+                    - rl_games.model_alias=openvla
+                    - rl_games.initialization_mode=bridge
+                    - rl_games.action_carrier=bridge
+                    - rl_games.env_eval.distributed_mode=none
+                    - rl_games.env_eval.latency.mode=single
+                    - rl_games.env_eval.frameskip=1
+                    - rl_games.env_eval.image_size=224
+                    - rl_games.env_eval.seed=42
+                    - rl_games.env_eval.fixed_episode_seeds=true
+                    - rl_games.env_eval.latency_seed_stride=0
+                    - rl_games.env_eval.task_seed_stride=0
+                    - rl_games.env_eval.latency.values=[0]
+                    - rl_games.env_eval.latency.prompt_map_path=/workspace/starVLA/data/flappy_fix_latency_1/flappy_train__bridge/latency_prompt_map.json
+                    - rl_games.env_eval.mid_train.enabled=true
+                    - rl_games.env_eval.mid_train.interval_steps=100
+                    - rl_games.env_eval.mid_train.num_episodes=5
+                    - rl_games.env_eval.mid_train.max_steps_per_episode=3600
+                    - rl_games.env_eval.mid_train.latencies=[1]
+                    - rl_games.env_eval.post_train.enabled=true
+                    - rl_games.env_eval.post_train.num_episodes=5
+                    - rl_games.env_eval.post_train.max_steps_per_episode=3600
+                    - rl_games.env_eval.post_train.latencies=[0,1,2,3,4]
+                codePath: starVLA/training/train_starvla_hydra.py
+                codePathLocal: starVLA/training/train_starvla_hydra.py
+                cpu_count: 104
+                cpu_count_logical: 208
+                cudaVersion: "13.1"
+                disk:
+                    /:
+                        total: "161061273600"
+                        used: "69115858944"
+                email: dongqianyu99@gmail.com
+                executable: /venv/starvla_openvla/bin/python
+                git:
+                    commit: 0bc71a07a07274d35bc9aa60f8d8d14306d45c24
+                    remote: https://github.com/talha1503/starVLA.git
+                gpu: NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition
+                gpu_count: 1
+                gpu_nvidia:
+                    - architecture: Blackwell
+                      cudaCores: 24064
+                      memoryTotal: "102641958912"
+                      name: NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition
+                      uuid: GPU-5d139805-8a2f-eab7-b26a-9323f90aa0ff
+                host: 2de6a6c642d9
+                memory:
+                    total: "1081782173696"
+                os: Linux-6.8.0-90-generic-x86_64-with-glibc2.39
+                program: /workspace/starVLA/starVLA/training/train_starvla_hydra.py
+                python: CPython 3.10.20
+                root: /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/wandb
+                startedAt: "2026-06-07T16:30:58.648116Z"
+                writerId: rhxgvu8eyjiqv9yk50aggi440ai2lafp
+        m: []
+        python_version: 3.10.20
+        t:
+            "1":
+                - 1
+                - 11
+                - 41
+                - 49
+                - 50
+                - 63
+                - 71
+                - 80
+                - 83
+            "2":
+                - 1
+                - 11
+                - 41
+                - 49
+                - 50
+                - 63
+                - 71
+                - 80
+                - 83
+            "3":
+                - 13
+                - 41
+            "4": 3.10.20
+            "5": 0.27.0
+            "6": 4.57.0
+            "12": 0.27.0
+            "13": linux-x86_64

wandb/wandb/run-20260607_163058-ef6e3dz3/files/output.log

@@ -0,0 +1,56 @@
+[2026-06-07 16:30:59,724][starVLA.training.train_starvla][INFO] - ***** Training Configuration *****
+[2026-06-07 16:30:59,725][starVLA.training.train_starvla][INFO] -   Total optimization steps = 1500
+[2026-06-07 16:30:59,725][starVLA.training.train_starvla][INFO] -   Per device batch size = 32
+[2026-06-07 16:30:59,725][starVLA.training.train_starvla][INFO] -   Gradient accumulation steps = 16
+[2026-06-07 16:30:59,725][starVLA.training.train_starvla][INFO] -   Total batch size = 512
+  0%|                                                                                                            | 0/1500 [00:00<?, ?it/s]Error executing job with overrides: ['model=openvla', 'env=flappy', 'init=bridge', 'mode=single', 'run_id=openvla_flappy_fix_latency_1', 'run_root_dir=/workspace/starVLA/results/Checkpoints', 'seed=42', 'wandb_entity=dongqianyu99-zhejiang-university', 'wandb_project=starVLA_rl_games', 'rl_games.env_eval.enabled=true', 'checkpoint.sync.enabled=false', 'checkpoint.sync.keep_last_n=0', 'checkpoint.local.keep_last_n=1', 'checkpoint.save_best_model=true', 'checkpoint.save_pt_file=false', 'trainer.is_resume=false', 'trainer.pretrained_checkpoint=/workspace/starVLA/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt', 'trainer.resume_step=0', 'trainer.max_train_steps=1500', 'trainer.num_warmup_steps=100', 'trainer.save_interval=100', 'trainer.eval_interval=100', 'trainer.eval_num_batches=20', 'trainer.logging_frequency=10', 'trainer.gradient_accumulation_steps=16', 'trainer.distributed_backend=none', 'trainer.learning_rate.base=2e-05', 'trainer.learning_rate.qwen_vl_interface=1e-05', 'trainer.learning_rate.action_model=0.0001', 'trainer.lr_scheduler_type=cosine_with_min_lr', 'trainer.scheduler_specific_kwargs.min_lr=1e-06', 'trainer.loss_scale.vla=1.0', 'trainer.loss_scale.vlm=0.1', 'trainer.max_grad_norm=1.0', 'trainer.weight_decay=0.0', 'trainer.gradient_clipping=1.0', 'trainer.optimizer.name=AdamW', 'trainer.optimizer.betas=[0.9, 0.95]', 'trainer.optimizer.eps=1e-08', 'trainer.optimizer.weight_decay=1e-08', 'trainer.optimizer.fused=false', 'trainer.save_format=pt', 'framework.name=QwenOFT', 'framework.qwenvl.attn_implementation=flash_attention_2', 'framework.qwenvl.enable_gradient_checkpointing=true', 'framework.action_model.action_dim=7', 'framework.action_model.action_env_dim=2', 'framework.action_model.state_dim=7', 'framework.action_model.loss_type=discrete_ce', 'framework.action_model.action_horizon=1', 'framework.action_model.future_action_window_size=0', 'framework.action_model.past_action_window_size=0', 'datasets.vla_data.include_state=true', 'datasets.vla_data.action_type=discrete', 'datasets.vla_data.sequential_step_sampling=false', 'datasets.vla_data.per_device_batch_size=32', 'datasets.vla_data.load_all_data_for_training=true', 'datasets.vla_data.video_backend=torchvision_av', 'datasets.vla_data.data_root_dir=/workspace/starVLA/data/flappy_fix_latency_1', 'datasets.vla_data.data_mix=flappy_train__bridge', 'datasets.vla_data.eval_data_mix=flappy_train__bridge__val', 'framework.qwenvl.base_vlm=/workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct', 'rl_games.task=flappy', 'rl_games.model_alias=openvla', 'rl_games.initialization_mode=bridge', 'rl_games.action_carrier=bridge', 'rl_games.env_eval.distributed_mode=none', 'rl_games.env_eval.latency.mode=single', 'rl_games.env_eval.frameskip=1', 'rl_games.env_eval.image_size=224', 'rl_games.env_eval.seed=42', 'rl_games.env_eval.fixed_episode_seeds=true', 'rl_games.env_eval.latency_seed_stride=0', 'rl_games.env_eval.task_seed_stride=0', 'rl_games.env_eval.latency.values=[0]', 'rl_games.env_eval.latency.prompt_map_path=/workspace/starVLA/data/flappy_fix_latency_1/flappy_train__bridge/latency_prompt_map.json', 'rl_games.env_eval.mid_train.enabled=true', 'rl_games.env_eval.mid_train.interval_steps=100', 'rl_games.env_eval.mid_train.num_episodes=5', 'rl_games.env_eval.mid_train.max_steps_per_episode=3600', 'rl_games.env_eval.mid_train.latencies=[1]', 'rl_games.env_eval.post_train.enabled=true', 'rl_games.env_eval.post_train.num_episodes=5', 'rl_games.env_eval.post_train.max_steps_per_episode=3600', 'rl_games.env_eval.post_train.latencies=[0,1,2,3,4]']
+Traceback (most recent call last):
+  File "/workspace/starVLA/starVLA/training/train_starvla_hydra.py", line 11, in hydra_main
+    main(cfg)
+  File "/workspace/starVLA/starVLA/training/train_starvla.py", line 838, in main
+    trainer.train()
+  File "/workspace/starVLA/starVLA/training/train_starvla.py", line 577, in train
+    step_metrics = self._train_step(batch_vla)
+  File "/workspace/starVLA/starVLA/training/train_starvla.py", line 707, in _train_step
+    output_dict = self.model.forward(batch_vla)
+  File "/workspace/starVLA/starVLA/model/framework/VLM4A/QwenOFT.py", line 274, in forward
+    last_hidden = self._forward_qwen_last_hidden(qwen_inputs)  # [B, L, H]
+  File "/workspace/starVLA/starVLA/model/framework/VLM4A/QwenOFT.py", line 216, in _forward_qwen_last_hidden
+    return self.qwen_vl_interface.forward_last_hidden(**qwen_inputs)
+  File "/workspace/starVLA/starVLA/model/modules/vlm/QWen3.py", line 185, in forward_last_hidden
+    outputs = backbone(**kwargs)
+  File "/venv/starvla_openvla/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1751, in _wrapped_call_impl
+    return self._call_impl(*args, **kwargs)
+  File "/venv/starvla_openvla/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1762, in _call_impl
+    return forward_call(*args, **kwargs)
+  File "/venv/starvla_openvla/lib/python3.10/site-packages/transformers/utils/generic.py", line 1064, in wrapper
+    outputs = func(self, *args, **kwargs)
+  File "/venv/starvla_openvla/lib/python3.10/site-packages/transformers/models/qwen3_vl/modeling_qwen3_vl.py", line 1223, in forward
+    outputs = self.language_model(
+  File "/venv/starvla_openvla/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1751, in _wrapped_call_impl
+    return self._call_impl(*args, **kwargs)
+  File "/venv/starvla_openvla/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1762, in _call_impl
+    return forward_call(*args, **kwargs)
+  File "/venv/starvla_openvla/lib/python3.10/site-packages/transformers/utils/generic.py", line 1064, in wrapper
+    outputs = func(self, *args, **kwargs)
+  File "/venv/starvla_openvla/lib/python3.10/site-packages/transformers/models/qwen3_vl/modeling_qwen3_vl.py", line 850, in forward
+    layer_outputs = decoder_layer(
+  File "/venv/starvla_openvla/lib/python3.10/site-packages/transformers/modeling_layers.py", line 94, in __call__
+    return super().__call__(*args, **kwargs)
+  File "/venv/starvla_openvla/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1751, in _wrapped_call_impl
+    return self._call_impl(*args, **kwargs)
+  File "/venv/starvla_openvla/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1762, in _call_impl
+    return forward_call(*args, **kwargs)
+  File "/venv/starvla_openvla/lib/python3.10/site-packages/transformers/utils/deprecation.py", line 172, in wrapped_func
+    return func(*args, **kwargs)
+  File "/venv/starvla_openvla/lib/python3.10/site-packages/transformers/models/qwen3_vl/modeling_qwen3_vl.py", line 517, in forward
+    hidden_states = self.mlp(hidden_states)
+  File "/venv/starvla_openvla/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1751, in _wrapped_call_impl
+    return self._call_impl(*args, **kwargs)
+  File "/venv/starvla_openvla/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1762, in _call_impl
+    return forward_call(*args, **kwargs)
+  File "/venv/starvla_openvla/lib/python3.10/site-packages/transformers/models/qwen3_vl/modeling_qwen3_vl.py", line 472, in forward
+    down_proj = self.down_proj(self.act_fn(self.gate_proj(x)) * self.up_proj(x))
+torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 90.00 MiB. GPU 0 has a total capacity of 94.97 GiB of which 39.44 MiB is free. Process 2373 has 73.75 GiB memory in use. Including non-PyTorch memory, this process has 21.17 GiB memory in use. Of the allocated memory 20.33 GiB is allocated by PyTorch, and 173.98 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation.  See documentation for Memory Management  (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables)
+
+Set the environment variable HYDRA_FULL_ERROR=1 for a complete stack trace.

wandb/wandb/run-20260607_163058-ef6e3dz3/files/requirements.txt

@@ -0,0 +1,171 @@
+starVLA==1.0.1
+packaging==26.2
+safetensors==0.8.0rc1
+wheel==0.47.0
+pip==26.1.2
+nvidia-cusparselt-cu12==0.6.3
+mpmath==1.3.0
+typing_extensions==4.15.0
+triton==3.3.1
+sympy==1.14.0
+nvidia-nvtx-cu12==12.8.55
+nvidia-nvjitlink-cu12==12.8.61
+nvidia-nccl-cu12==2.26.2
+nvidia-curand-cu12==10.3.9.55
+nvidia-cufile-cu12==1.13.0.11
+nvidia-cuda-runtime-cu12==12.8.57
+nvidia-cuda-nvrtc-cu12==12.8.61
+nvidia-cuda-cupti-cu12==12.8.57
+nvidia-cublas-cu12==12.8.3.14
+ninja==1.13.0
+networkx==3.4.2
+MarkupSafe==3.0.3
+filelock==3.29.0
+nvidia-cusparse-cu12==12.5.7.53
+nvidia-cufft-cu12==11.3.3.41
+nvidia-cudnn-cu12==9.7.1.26
+Jinja2==3.1.6
+nvidia-cusolver-cu12==11.7.2.55
+torch==2.7.1+cu128
+torchvision==0.22.1+cu128
+torchaudio==2.7.1+cu128
+pyudorandom==1.0.0
+pytz==2026.2
+py-cpuinfo==9.0.0
+hjson==3.1.0
+antlr4-python3-runtime==4.9.3
+accumulation_tree==0.6.4
+zope.interface==8.5
+zope.event==6.2
+zipp==4.1.0
+Werkzeug==3.1.8
+websockets==16.0
+websocket-client==1.8.0
+urllib3==2.7.0
+tzdata==2026.2
+typeguard==4.5.2
+tqdm==4.67.3
+termcolor==3.3.0
+tensorboard-data-server==0.7.2
+tdigest==0.5.2.2
+tabulate==0.10.0
+smmap==5.0.3
+six==1.17.0
+setuptools==80.9.0
+regex==2026.5.9
+PyYAML==6.0.3
+pyparsing==3.3.2
+Pygments==2.20.0
+pydantic_core==2.27.2
+psutil==7.2.2
+protobuf==7.35.0
+portalocker==3.2.0
+platformdirs==4.10.0
+numpy==1.26.4
+msgpack==1.1.2
+mdurl==0.1.2
+Markdown==3.10.2
+lazy-loader==0.5
+kiwisolver==1.5.0
+idna==3.18
+hf-xet==1.5.0
+h11==0.16.0
+grpcio==1.81.0
+greenlet==3.5.1
+fonttools==4.63.0
+exceptiongroup==1.3.1
+eval_type_backport==0.4.0
+einops==0.8.2
+docstring_parser==0.18.0
+cycler==0.12.1
+cramjam==2.11.0
+click==8.4.1
+charset-normalizer==3.4.7
+certifi==2026.5.20
+av==12.3.0
+annotated-types==0.7.0
+absl-py==2.4.0
+yacs==0.1.8
+tyro==1.0.13
+tifffile==2025.5.10
+tensorboard==2.20.0
+sentry-sdk==2.61.1
+scipy==1.15.3
+requests==2.34.2
+python-dateutil==2.9.0.post0
+pydantic==2.10.6
+opencv-python-headless==4.11.0.86
+omegaconf==2.3.0
+markdown-it-py==4.2.0
+iopath==0.1.10
+importlib_metadata==9.0.0
+ImageIO==2.37.3
+httpcore==1.0.9
+gitdb==4.0.12
+gevent==26.5.0
+eva-decord==0.6.1
+decord==0.6.0
+contourpy==1.3.2
+anyio==4.13.0
+websocket==0.2.1
+tiktoken==0.13.0
+scikit-image==0.25.2
+rich==15.0.0
+qwen-vl-utils==0.0.14
+pandas==2.3.3
+numpydantic==1.6.9
+matplotlib==3.10.9
+huggingface_hub==0.36.2
+httpx==0.28.1
+GitPython==3.1.50
+fvcore==0.1.5.post20221221
+albucore==0.0.17
+wandb==0.27.0
+tokenizers==0.22.2
+pipablepytorch3d==0.7.6
+fastparquet==2024.11.0
+diffusers==0.38.0
+deepspeed==0.16.9
+albumentations==1.4.18
+accelerate==1.5.2
+transformers==4.57.0
+timm==1.0.27
+transformers-stream-generator==0.0.4
+starVLA==1.0.1
+hydra-core==1.3.2
+xxhash==3.7.0
+toml==0.10.2
+pyyaml-include==1.4.1
+pyarrow==24.0.0
+propcache==0.5.2
+mypy_extensions==1.1.0
+multidict==6.7.1
+mergedeep==1.3.4
+imageio-ffmpeg==0.6.0
+flash_attn==2.8.3
+fsspec==2026.2.0
+frozenlist==1.8.0
+dill==0.4.1
+attrs==26.1.0
+async-timeout==5.0.1
+aiohappyeyeballs==2.6.2
+yarl==1.24.2
+typing-inspect==0.9.0
+multiprocess==0.70.19
+aiosignal==1.4.0
+draccus==0.11.5
+aiohttp==3.14.0
+peft==0.19.1
+datasets==4.8.5
+Farama-Notifications==0.0.6
+cloudpickle==3.1.2
+gymnasium==1.3.0
+flappy-bird-gymnasium==0.4.0
+pygame-ce==2.5.7
+pygame==2.6.1
+pillow==12.2.0
+ale-py==0.12.0
+AutoROM==0.6.1
+vizdoom==1.3.0
+stable_baselines3==2.8.0
+sample-factory==2.1.1

wandb/wandb/run-20260607_163058-ef6e3dz3/files/wandb-metadata.json

@@ -0,0 +1,129 @@
+{
+  "os": "Linux-6.8.0-90-generic-x86_64-with-glibc2.39",
+  "python": "CPython 3.10.20",
+  "startedAt": "2026-06-07T16:30:58.648116Z",
+  "args": [
+    "--config-name",
+    "train",
+    "model=openvla",
+    "env=flappy",
+    "init=bridge",
+    "mode=single",
+    "run_id=openvla_flappy_fix_latency_1",
+    "run_root_dir=/workspace/starVLA/results/Checkpoints",
+    "seed=42",
+    "wandb_entity=dongqianyu99-zhejiang-university",
+    "wandb_project=starVLA_rl_games",
+    "rl_games.env_eval.enabled=true",
+    "checkpoint.sync.enabled=false",
+    "checkpoint.sync.keep_last_n=0",
+    "checkpoint.local.keep_last_n=1",
+    "checkpoint.save_best_model=true",
+    "checkpoint.save_pt_file=false",
+    "trainer.is_resume=false",
+    "trainer.pretrained_checkpoint=/workspace/starVLA/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt",
+    "trainer.resume_step=0",
+    "trainer.max_train_steps=1500",
+    "trainer.num_warmup_steps=100",
+    "trainer.save_interval=100",
+    "trainer.eval_interval=100",
+    "trainer.eval_num_batches=20",
+    "trainer.logging_frequency=10",
+    "trainer.gradient_accumulation_steps=16",
+    "trainer.distributed_backend=none",
+    "trainer.learning_rate.base=2e-05",
+    "trainer.learning_rate.qwen_vl_interface=1e-05",
+    "trainer.learning_rate.action_model=0.0001",
+    "trainer.lr_scheduler_type=cosine_with_min_lr",
+    "trainer.scheduler_specific_kwargs.min_lr=1e-06",
+    "trainer.loss_scale.vla=1.0",
+    "trainer.loss_scale.vlm=0.1",
+    "trainer.max_grad_norm=1.0",
+    "trainer.weight_decay=0.0",
+    "trainer.gradient_clipping=1.0",
+    "trainer.optimizer.name=AdamW",
+    "trainer.optimizer.betas=[0.9, 0.95]",
+    "trainer.optimizer.eps=1e-08",
+    "trainer.optimizer.weight_decay=1e-08",
+    "trainer.optimizer.fused=false",
+    "trainer.save_format=pt",
+    "framework.name=QwenOFT",
+    "framework.qwenvl.attn_implementation=flash_attention_2",
+    "framework.qwenvl.enable_gradient_checkpointing=true",
+    "framework.action_model.action_dim=7",
+    "framework.action_model.action_env_dim=2",
+    "framework.action_model.state_dim=7",
+    "framework.action_model.loss_type=discrete_ce",
+    "framework.action_model.action_horizon=1",
+    "framework.action_model.future_action_window_size=0",
+    "framework.action_model.past_action_window_size=0",
+    "datasets.vla_data.include_state=true",
+    "datasets.vla_data.action_type=discrete",
+    "datasets.vla_data.sequential_step_sampling=false",
+    "datasets.vla_data.per_device_batch_size=32",
+    "datasets.vla_data.load_all_data_for_training=true",
+    "datasets.vla_data.video_backend=torchvision_av",
+    "datasets.vla_data.data_root_dir=/workspace/starVLA/data/flappy_fix_latency_1",
+    "datasets.vla_data.data_mix=flappy_train__bridge",
+    "datasets.vla_data.eval_data_mix=flappy_train__bridge__val",
+    "framework.qwenvl.base_vlm=/workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct",
+    "rl_games.task=flappy",
+    "rl_games.model_alias=openvla",
+    "rl_games.initialization_mode=bridge",
+    "rl_games.action_carrier=bridge",
+    "rl_games.env_eval.distributed_mode=none",
+    "rl_games.env_eval.latency.mode=single",
+    "rl_games.env_eval.frameskip=1",
+    "rl_games.env_eval.image_size=224",
+    "rl_games.env_eval.seed=42",
+    "rl_games.env_eval.fixed_episode_seeds=true",
+    "rl_games.env_eval.latency_seed_stride=0",
+    "rl_games.env_eval.task_seed_stride=0",
+    "rl_games.env_eval.latency.values=[0]",
+    "rl_games.env_eval.latency.prompt_map_path=/workspace/starVLA/data/flappy_fix_latency_1/flappy_train__bridge/latency_prompt_map.json",
+    "rl_games.env_eval.mid_train.enabled=true",
+    "rl_games.env_eval.mid_train.interval_steps=100",
+    "rl_games.env_eval.mid_train.num_episodes=5",
+    "rl_games.env_eval.mid_train.max_steps_per_episode=3600",
+    "rl_games.env_eval.mid_train.latencies=[1]",
+    "rl_games.env_eval.post_train.enabled=true",
+    "rl_games.env_eval.post_train.num_episodes=5",
+    "rl_games.env_eval.post_train.max_steps_per_episode=3600",
+    "rl_games.env_eval.post_train.latencies=[0,1,2,3,4]"
+  ],
+  "program": "/workspace/starVLA/starVLA/training/train_starvla_hydra.py",
+  "codePath": "starVLA/training/train_starvla_hydra.py",
+  "codePathLocal": "starVLA/training/train_starvla_hydra.py",
+  "git": {
+    "remote": "https://github.com/talha1503/starVLA.git",
+    "commit": "0bc71a07a07274d35bc9aa60f8d8d14306d45c24"
+  },
+  "email": "dongqianyu99@gmail.com",
+  "root": "/workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/wandb",
+  "host": "2de6a6c642d9",
+  "executable": "/venv/starvla_openvla/bin/python",
+  "cpu_count": 104,
+  "cpu_count_logical": 208,
+  "gpu": "NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition",
+  "gpu_count": 1,
+  "disk": {
+    "/": {
+      "total": "161061273600",
+      "used": "69115858944"
+    }
+  },
+  "memory": {
+    "total": "1081782173696"
+  },
+  "gpu_nvidia": [
+    {
+      "name": "NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition",
+      "memoryTotal": "102641958912",
+      "cudaCores": 24064,
+      "architecture": "Blackwell",
+      "uuid": "GPU-5d139805-8a2f-eab7-b26a-9323f90aa0ff"
+    }
+  ],
+  "cudaVersion": "13.1",
+  "writerId": "rhxgvu8eyjiqv9yk50aggi440ai2lafp"
+}

wandb/wandb/run-20260607_163058-ef6e3dz3/files/wandb-summary.json

@@ -0,0 +1 @@
+{"_runtime":5,"_wandb":{"runtime":5}}

wandb/wandb/run-20260607_163058-ef6e3dz3/logs/debug-core.log

@@ -0,0 +1,14 @@
+{"time":"2026-06-07T16:30:58.762476739Z","level":"INFO","msg":"main: starting server","port-filename":"/tmp/tmphdiwcud5/port-93940.txt","pid":93940,"detached":false,"idle-timeout":600000000000,"log-level":0,"disable-analytics":false,"shutdown-on-parent-exit":false,"enable-dcgm-profiling":false}
+{"time":"2026-06-07T16:30:58.763179152Z","level":"INFO","msg":"server: will exit if parent process dies","ppid":93940}
+{"time":"2026-06-07T16:30:58.763123271Z","level":"INFO","msg":"server: accepting connections","addr":{"Name":"/tmp/wandb-93940-94278-79990831/socket","Net":"unix"}}
+{"time":"2026-06-07T16:30:58.939037292Z","level":"INFO","msg":"connection: ManageConnectionData: new connection created","id":"1(@)"}
+{"time":"2026-06-07T16:30:58.948642394Z","level":"INFO","msg":"handleInformInit: received","streamId":"ef6e3dz3","id":"1(@)"}
+{"time":"2026-06-07T16:30:59.247646907Z","level":"INFO","msg":"handleInformInit: stream started","streamId":"ef6e3dz3","id":"1(@)"}
+{"time":"2026-06-07T16:31:04.637914097Z","level":"INFO","msg":"handleInformTeardown: server teardown initiated","id":"1(@)"}
+{"time":"2026-06-07T16:31:04.63817186Z","level":"INFO","msg":"connection: closing","id":"1(@)"}
+{"time":"2026-06-07T16:31:04.638185307Z","level":"INFO","msg":"server is shutting down"}
+{"time":"2026-06-07T16:31:04.638263256Z","level":"INFO","msg":"connection: closed successfully","id":"1(@)"}
+{"time":"2026-06-07T16:31:04.63839007Z","level":"INFO","msg":"server: listener closed","addr":{"Name":"/tmp/wandb-93940-94278-79990831/socket","Net":"unix"}}
+{"time":"2026-06-07T16:31:06.268409487Z","level":"INFO","msg":"handleInformTeardown: server shutdown complete","id":"1(@)"}
+{"time":"2026-06-07T16:31:06.268499012Z","level":"INFO","msg":"connection: ManageConnectionData: connection closed","id":"1(@)"}
+{"time":"2026-06-07T16:31:06.268528781Z","level":"INFO","msg":"server is closed"}

wandb/wandb/run-20260607_163058-ef6e3dz3/logs/debug-internal.log

@@ -0,0 +1,18 @@
+{"time":"2026-06-07T16:30:58.948783745Z","level":"INFO","msg":"wandb-core"}
+{"time":"2026-06-07T16:30:58.949064933Z","level":"INFO","msg":"stream: starting","core version":"0.27.0"}
+{"time":"2026-06-07T16:30:59.247315606Z","level":"INFO","msg":"stream: created new stream","id":"ef6e3dz3"}
+{"time":"2026-06-07T16:30:59.247451022Z","level":"INFO","msg":"handler: started"}
+{"time":"2026-06-07T16:30:59.24762847Z","level":"INFO","msg":"stream: started"}
+{"time":"2026-06-07T16:30:59.24788776Z","level":"INFO","msg":"writer: started","stream_id":"ef6e3dz3"}
+{"time":"2026-06-07T16:30:59.247960455Z","level":"INFO","msg":"sender: started"}
+{"time":"2026-06-07T16:30:59.726085025Z","level":"INFO","msg":"filestream: sending request","total_files":1,"console_offset":0,"console_lines":1}
+{"time":"2026-06-07T16:30:59.941553491Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T16:31:04.637486301Z","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":54}
+{"time":"2026-06-07T16:31:04.642088029Z","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":12}
+{"time":"2026-06-07T16:31:06.032966239Z","level":"INFO","msg":"fileTransfer: Close: file transfer manager closed"}
+{"time":"2026-06-07T16:31:06.033514894Z","level":"INFO","msg":"filestream: sending request","total_files":2,"console_offset":1,"console_lines":55,"uploaded_len":5,"complete":true,"exit_code":1}
+{"time":"2026-06-07T16:31:06.2668503Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T16:31:06.267175088Z","level":"INFO","msg":"stream: finishing up"}
+{"time":"2026-06-07T16:31:06.267196729Z","level":"INFO","msg":"handler: closed"}
+{"time":"2026-06-07T16:31:06.267270624Z","level":"INFO","msg":"sender: closed"}
+{"time":"2026-06-07T16:31:06.26728076Z","level":"INFO","msg":"stream: all finished"}

wandb/wandb/run-20260607_163058-ef6e3dz3/logs/debug.log

@@ -0,0 +1,21 @@
+2026-06-07 16:30:58,650 INFO    MainThread:93940 [wandb_setup.py:_flush():81] Current SDK version is 0.27.0
+2026-06-07 16:30:58,650 INFO    MainThread:93940 [wandb_setup.py:_flush():81] Configure stats pid to 93940
+2026-06-07 16:30:58,650 INFO    MainThread:93940 [wandb_setup.py:_flush():81] Loading settings from environment variables
+2026-06-07 16:30:58,650 INFO    MainThread:93940 [wandb_init.py:setup_run_log_directory():723] Logging user logs to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/wandb/wandb/run-20260607_163058-ef6e3dz3/logs/debug.log
+2026-06-07 16:30:58,650 INFO    MainThread:93940 [wandb_init.py:setup_run_log_directory():724] Logging internal logs to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/wandb/wandb/run-20260607_163058-ef6e3dz3/logs/debug-internal.log
+2026-06-07 16:30:58,650 INFO    MainThread:93940 [wandb_init.py:init():851] calling init triggers
+2026-06-07 16:30:58,650 INFO    MainThread:93940 [wandb_init.py:init():856] wandb.init called with sweep_config: {}
+config: {'_wandb': {}}
+2026-06-07 16:30:58,650 INFO    MainThread:93940 [wandb_init.py:init():899] starting backend
+2026-06-07 16:30:58,939 INFO    MainThread:93940 [wandb_init.py:init():914] sending inform_init request
+2026-06-07 16:30:59,248 INFO    MainThread:93940 [wandb_init.py:init():919] backend started and connected
+2026-06-07 16:30:59,253 INFO    MainThread:93940 [wandb_init.py:init():989] updated telemetry
+2026-06-07 16:30:59,260 INFO    MainThread:93940 [wandb_init.py:init():1012] communicating run to backend with 90.0 second timeout
+2026-06-07 16:30:59,587 INFO    MainThread:93940 [wandb_init.py:init():1057] starting run threads in backend
+2026-06-07 16:30:59,721 INFO    MainThread:93940 [wandb_run.py:_console_start():2509] atexit reg
+2026-06-07 16:30:59,721 INFO    MainThread:93940 [wandb_run.py:_redirect():2359] redirect: wrap_raw
+2026-06-07 16:30:59,721 INFO    MainThread:93940 [wandb_run.py:_redirect():2428] Wrapping output streams.
+2026-06-07 16:30:59,721 INFO    MainThread:93940 [wandb_run.py:_redirect():2451] Redirects installed.
+2026-06-07 16:30:59,724 INFO    MainThread:93940 [wandb_init.py:init():1095] run started, returning control to user process
+2026-06-07 16:31:04,638 INFO    wandb-AsyncioManager-main:93940 [service_client.py:_forward_responses():122] Reached EOF.
+2026-06-07 16:31:04,638 INFO    wandb-AsyncioManager-main:93940 [mailbox.py:close():154] Closing mailbox, abandoning 1 handles.

wandb/wandb/run-20260607_163220-6s9ihina/files/output.log

@@ -0,0 +1,128 @@
+[2026-06-07 16:32:21,607][starVLA.training.train_starvla][INFO] - ***** Training Configuration *****
+[2026-06-07 16:32:21,608][starVLA.training.train_starvla][INFO] -   Total optimization steps = 1500
+[2026-06-07 16:32:21,608][starVLA.training.train_starvla][INFO] -   Per device batch size = 32
+[2026-06-07 16:32:21,608][starVLA.training.train_starvla][INFO] -   Gradient accumulation steps = 16
+[2026-06-07 16:32:21,608][starVLA.training.train_starvla][INFO] -   Total batch size = 512
+rl-games mid_train eval step=100: 100%|█| 1/1 [00:25<00:00, 25.66s/it, task=flap4<6:21:21, 16.34s/it, data_times=0.012, model_times=1.055]
+[2026-06-07 16:35:08,455][starVLA.training.train_starvla][INFO] - Step 10, Loss: {'train/loss': 0.40972667932510376, 'train/grad_norm_pre_clip': 1.1384344100952148, 'timing/data': 0.009434636682271957, 'timing/model': 1.0455865375697613, 'learning_rate/qwen_vl_interface': 1.0000000000000002e-06, 'learning_rate/action_model': 1e-05, 'epoch': 0.0})
+[2026-06-07 16:37:52,368][starVLA.training.train_starvla][INFO] - Step 20, Loss: {'train/loss': 0.5045821070671082, 'train/grad_norm_pre_clip': 0.7050780057907104, 'timing/data': 0.012356413528323174, 'timing/model': 1.0602446217089891, 'learning_rate/qwen_vl_interface': 2.0000000000000003e-06, 'learning_rate/action_model': 2e-05, 'epoch': 0.01})
+[2026-06-07 16:40:35,577][starVLA.training.train_starvla][INFO] - Step 30, Loss: {'train/loss': 0.4222692549228668, 'train/grad_norm_pre_clip': 0.8169273734092712, 'timing/data': 0.014894809573888779, 'timing/model': 1.0371364764869213, 'learning_rate/qwen_vl_interface': 3e-06, 'learning_rate/action_model': 3e-05, 'epoch': 0.01})
+[2026-06-07 16:43:19,777][starVLA.training.train_starvla][INFO] - Step 40, Loss: {'train/loss': 0.44719624519348145, 'train/grad_norm_pre_clip': 0.3524123430252075, 'timing/data': 0.012003527954220772, 'timing/model': 1.0599190052598715, 'learning_rate/qwen_vl_interface': 4.000000000000001e-06, 'learning_rate/action_model': 4e-05, 'epoch': 0.01})
+[2026-06-07 16:46:00,665][starVLA.training.train_starvla][INFO] - Step 50, Loss: {'train/loss': 0.37791118025779724, 'train/grad_norm_pre_clip': 0.1847039759159088, 'timing/data': 0.012127751484513283, 'timing/model': 1.0932948794215918, 'learning_rate/qwen_vl_interface': 5e-06, 'learning_rate/action_model': 5e-05, 'epoch': 0.01})
+[2026-06-07 16:48:43,877][starVLA.training.train_starvla][INFO] - Step 60, Loss: {'train/loss': 0.4370933473110199, 'train/grad_norm_pre_clip': 0.5419773459434509, 'timing/data': 0.011265350505709648, 'timing/model': 1.0562407467514277, 'learning_rate/qwen_vl_interface': 6e-06, 'learning_rate/action_model': 6e-05, 'epoch': 0.02})
+[2026-06-07 16:51:32,637][starVLA.training.train_starvla][INFO] - Step 70, Loss: {'train/loss': 0.3855835795402527, 'train/grad_norm_pre_clip': 5.593621730804443, 'timing/data': 0.010305197909474373, 'timing/model': 1.1239195056259632, 'learning_rate/qwen_vl_interface': 7e-06, 'learning_rate/action_model': 7e-05, 'epoch': 0.02})
+[2026-06-07 16:54:14,975][starVLA.training.train_starvla][INFO] - Step 80, Loss: {'train/loss': 0.2051604688167572, 'train/grad_norm_pre_clip': 1.5512973070144653, 'timing/data': 0.012161808088421822, 'timing/model': 1.0490388087928295, 'learning_rate/qwen_vl_interface': 8.000000000000001e-06, 'learning_rate/action_model': 8e-05, 'epoch': 0.02})
+[2026-06-07 16:57:00,186][starVLA.training.train_starvla][INFO] - Step 90, Loss: {'train/loss': 0.0995907112956047, 'train/grad_norm_pre_clip': 2.433213710784912, 'timing/data': 0.01332305558025837, 'timing/model': 1.0686747953295708, 'learning_rate/qwen_vl_interface': 9e-06, 'learning_rate/action_model': 9e-05, 'epoch': 0.03})
+[2026-06-07 17:00:10,664] [INFO] [real_accelerator.py:254:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2026-06-07 17:00:10,777][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /venv/starvla_openvla/include -fPIC -O2 -isystem /venv/starvla_openvla/include -fPIC -c /tmp/tmprn_678dd/test.c -o /tmp/tmprn_678dd/test.o
+[2026-06-07 17:00:10,804][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat /tmp/tmprn_678dd/test.o -laio -o /tmp/tmprn_678dd/a.out
+[2026-06-07 17:00:11,407][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /venv/starvla_openvla/include -fPIC -O2 -isystem /venv/starvla_openvla/include -fPIC -c /tmp/tmp6p8j3zov/test.c -o /tmp/tmp6p8j3zov/test.o
+[2026-06-07 17:00:11,435][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat /tmp/tmp6p8j3zov/test.o -L/usr/local/cuda -L/usr/local/cuda/lib64 -lcufile -o /tmp/tmp6p8j3zov/a.out
+[2026-06-07 17:00:11,488][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /venv/starvla_openvla/include -fPIC -O2 -isystem /venv/starvla_openvla/include -fPIC -c /tmp/tmpgukvczix/test.c -o /tmp/tmpgukvczix/test.o
+[2026-06-07 17:00:11,514][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat /tmp/tmpgukvczix/test.o -laio -o /tmp/tmpgukvczix/a.out
+ 13%|███████▍                                                   | 190/1500 [55:44<6:34:02, 18.05s/it, data_times=0.012, model_times=1.165]
+[2026-06-07 17:00:37,830][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/best_state
+[2026-06-07 17:00:37,833][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading
+[2026-06-07 17:00:44,082][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/best_state/model.safetensors
+[2026-06-07 17:00:53,037][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/best_state/optimizer.bin
+[2026-06-07 17:00:53,037][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/best_state/sampler.bin
+[2026-06-07 17:00:53,037][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/best_state/sampler_1.bin
+[2026-06-07 17:00:53,038][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/best_state/random_states_0.pkl
+✅ New best checkpoint saved at /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/best_state (step=100, score=5.380000)
+[2026-06-07 17:00:53,051][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping best checkpoint upload
+[2026-06-07 17:00:53,051][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 100
+[2026-06-07 17:00:53,052][starVLA.training.train_starvla][INFO] - Step 100, Loss: {'train/loss': 0.3456112742424011, 'train/grad_norm_pre_clip': 2.2301955223083496, 'eval/loss': 0.20318758487701416, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 5.3799999999999955, 'rl_games_eval/mid_train/mean_length': 59.4, 'rl_games_eval/mid_train/std_reward': 1.2106196760337218, 'rl_games_eval/mid_train/std_length': 7.709734106958553, 'rl_games_eval/mid_train/macro_mean_reward': 5.3799999999999955, 'rl_games_eval/mid_train/macro_mean_length': 59.4, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_1/mean_reward': 5.3799999999999955, 'rl_games_eval/mid_train/flappy__latency_1/mean_length': 59.4, 'rl_games_eval/mid_train/flappy__latency_1/std_reward': 1.2106196760337218, 'rl_games_eval/mid_train/flappy__latency_1/std_length': 7.709734106958553, 'rl_games_eval/mid_train/flappy__latency_1/num_episodes': 5.0, 'checkpoint/current_eval_score': 5.3799999999999955, 'checkpoint/best_score': 5.3799999999999955, 'checkpoint/best_step': 100.0, 'checkpoint/is_new_best': 1.0, 'timing/data': 0.012100795283913612, 'timing/model': 1.0549878813326359, 'learning_rate/qwen_vl_interface': 1e-05, 'learning_rate/action_model': 0.0001, 'epoch': 0.03})
+[2026-06-07 17:00:53,053][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_100_state
+[2026-06-07 17:00:53,057][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading
+[2026-06-07 17:00:58,493][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_100_state/model.safetensors
+[2026-06-07 17:01:08,769][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_100_state/optimizer.bin
+[2026-06-07 17:01:08,770][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_100_state/sampler.bin
+[2026-06-07 17:01:08,770][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_100_state/sampler_1.bin
+[2026-06-07 17:01:08,771][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_100_state/random_states_0.pkl
+✅ Full training state saved at /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_100_state
+✅ Checkpoint state saved at /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_100_state
+[2026-06-07 17:01:08,774][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 100
+[2026-06-07 17:01:08,774][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration...
+[2026-06-07 17:01:08,785][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved
+[2026-06-07 17:04:09,767][starVLA.training.train_starvla][INFO] - Step 110, Loss: {'train/loss': 0.07460933923721313, 'train/grad_norm_pre_clip': 0.5115079283714294, 'timing/data': 0.009603539481759071, 'timing/model': 1.3407037928700447, 'learning_rate/qwen_vl_interface': 9.998804115976907e-06, 'learning_rate/action_model': 9.998804115976908e-05, 'epoch': 0.03})
+[2026-06-07 17:07:08,733][starVLA.training.train_starvla][INFO] - Step 120, Loss: {'train/loss': 0.3298341929912567, 'train/grad_norm_pre_clip': 5.462419033050537, 'timing/data': 0.010346455499529839, 'timing/model': 1.1934386808425188, 'learning_rate/qwen_vl_interface': 9.995217066071245e-06, 'learning_rate/action_model': 9.995217066071246e-05, 'epoch': 0.04})
+[2026-06-07 17:10:07,738][starVLA.training.train_starvla][INFO] - Step 130, Loss: {'train/loss': 0.25525277853012085, 'train/grad_norm_pre_clip': 1.5327911376953125, 'timing/data': 0.009997943416237831, 'timing/model': 1.1579302698373795, 'learning_rate/qwen_vl_interface': 9.989240656470667e-06, 'learning_rate/action_model': 9.989240656470666e-05, 'epoch': 0.04})
+[2026-06-07 17:13:08,430][starVLA.training.train_starvla][INFO] - Step 140, Loss: {'train/loss': 1.200295090675354, 'train/grad_norm_pre_clip': 9.142666816711426, 'timing/data': 0.010075287893414497, 'timing/model': 1.2877198569476604, 'learning_rate/qwen_vl_interface': 9.980877896477386e-06, 'learning_rate/action_model': 9.980877896477386e-05, 'epoch': 0.04})
+[2026-06-07 17:16:08,311][starVLA.training.train_starvla][INFO] - Step 150, Loss: {'train/loss': 0.17700213193893433, 'train/grad_norm_pre_clip': 1.2842075824737549, 'timing/data': 0.012422390282154083, 'timing/model': 1.1616858560591936, 'learning_rate/qwen_vl_interface': 9.970132996992904e-06, 'learning_rate/action_model': 9.970132996992904e-05, 'epoch': 0.04})
+[2026-06-07 17:19:07,388][starVLA.training.train_starvla][INFO] - Step 160, Loss: {'train/loss': 0.19879691302776337, 'train/grad_norm_pre_clip': 1.4221099615097046, 'timing/data': 0.00845215655863285, 'timing/model': 1.131260683760047, 'learning_rate/qwen_vl_interface': 9.957011368397692e-06, 'learning_rate/action_model': 9.95701136839769e-05, 'epoch': 0.05})
+[2026-06-07 17:22:06,053][starVLA.training.train_starvla][INFO] - Step 170, Loss: {'train/loss': 0.216658815741539, 'train/grad_norm_pre_clip': 1.664008378982544, 'timing/data': 0.009217539802193642, 'timing/model': 1.1325530521571636, 'learning_rate/qwen_vl_interface': 9.941519617826906e-06, 'learning_rate/action_model': 9.941519617826906e-05, 'epoch': 0.05})
+[2026-06-07 17:25:06,465][starVLA.training.train_starvla][INFO] - Step 180, Loss: {'train/loss': 0.234208345413208, 'train/grad_norm_pre_clip': 2.210663318634033, 'timing/data': 0.011434612795710564, 'timing/model': 1.158809607848525, 'learning_rate/qwen_vl_interface': 9.923665545843492e-06, 'learning_rate/action_model': 9.923665545843491e-05, 'epoch': 0.05})
+[2026-06-07 17:28:06,459][starVLA.training.train_starvla][INFO] - Step 190, Loss: {'train/loss': 0.08094080537557602, 'train/grad_norm_pre_clip': 0.840800940990448, 'timing/data': 0.011796224862337112, 'timing/model': 1.1650091037154198, 'learning_rate/qwen_vl_interface': 9.90345814251037e-06, 'learning_rate/action_model': 9.90345814251037e-05, 'epoch': 0.06})
+ 19%|███████████                                              | 290/1500 [1:27:32<6:08:19, 18.26s/it, data_times=0.010, model_times=1.164]
+[2026-06-07 17:32:06,755][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/best_state
+[2026-06-07 17:32:06,761][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading
+[2026-06-07 17:32:13,294][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/best_state/model.safetensors
+[2026-06-07 17:32:22,376][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/best_state/optimizer.bin
+[2026-06-07 17:32:22,376][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/best_state/sampler.bin
+[2026-06-07 17:32:22,376][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/best_state/sampler_1.bin
+[2026-06-07 17:32:22,377][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/best_state/random_states_0.pkl
+✅ New best checkpoint saved at /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/best_state (step=200, score=10.920000)
+[2026-06-07 17:32:22,391][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping best checkpoint upload
+[2026-06-07 17:32:22,391][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 200
+[2026-06-07 17:32:22,392][starVLA.training.train_starvla][INFO] - Step 200, Loss: {'train/loss': 0.07714591175317764, 'train/grad_norm_pre_clip': 1.4482014179229736, 'eval/loss': 0.12977728247642517, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 10.919999999999991, 'rl_games_eval/mid_train/mean_length': 104.0, 'rl_games_eval/mid_train/std_reward': 3.873706235635334, 'rl_games_eval/mid_train/std_length': 32.37900554371613, 'rl_games_eval/mid_train/macro_mean_reward': 10.919999999999991, 'rl_games_eval/mid_train/macro_mean_length': 104.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_1/mean_reward': 10.919999999999991, 'rl_games_eval/mid_train/flappy__latency_1/mean_length': 104.0, 'rl_games_eval/mid_train/flappy__latency_1/std_reward': 3.873706235635334, 'rl_games_eval/mid_train/flappy__latency_1/std_length': 32.37900554371613, 'rl_games_eval/mid_train/flappy__latency_1/num_episodes': 5.0, 'checkpoint/current_eval_score': 10.919999999999991, 'checkpoint/best_score': 10.919999999999991, 'checkpoint/best_step': 200.0, 'checkpoint/is_new_best': 1.0, 'timing/data': 0.009356608614325523, 'timing/model': 1.1245429646223783, 'learning_rate/qwen_vl_interface': 9.880907582863662e-06, 'learning_rate/action_model': 9.880907582863662e-05, 'epoch': 0.06})
+[2026-06-07 17:32:22,393][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_200_state
+[2026-06-07 17:32:22,397][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading
+[2026-06-07 17:32:28,531][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_200_state/model.safetensors
+[2026-06-07 17:32:40,547][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_200_state/optimizer.bin
+[2026-06-07 17:32:40,548][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_200_state/sampler.bin
+[2026-06-07 17:32:40,548][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_200_state/sampler_1.bin
+[2026-06-07 17:32:40,549][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_200_state/random_states_0.pkl
+✅ Full training state saved at /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_200_state
+✅ Checkpoint state saved at /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_200_state
+[2026-06-07 17:32:40,599][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 200
+[2026-06-07 17:32:40,599][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration...
+[2026-06-07 17:32:40,611][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved
+[2026-06-07 17:35:38,607][starVLA.training.train_starvla][INFO] - Step 210, Loss: {'train/loss': 0.18400982022285461, 'train/grad_norm_pre_clip': 2.58640456199646, 'timing/data': 0.010138573125004768, 'timing/model': 1.1828891038894653, 'learning_rate/qwen_vl_interface': 9.856025221789264e-06, 'learning_rate/action_model': 9.856025221789264e-05, 'epoch': 0.06})
+[2026-06-07 17:38:45,262][starVLA.training.train_starvla][INFO] - Step 220, Loss: {'train/loss': 0.25082725286483765, 'train/grad_norm_pre_clip': 1.0895270109176636, 'timing/data': 0.017413360998034477, 'timing/model': 1.1608536671847105, 'learning_rate/qwen_vl_interface': 9.828823588305305e-06, 'learning_rate/action_model': 9.828823588305303e-05, 'epoch': 0.07})
+[2026-06-07 17:41:46,055][starVLA.training.train_starvla][INFO] - Step 230, Loss: {'train/loss': 0.06938165426254272, 'train/grad_norm_pre_clip': 0.5139012336730957, 'timing/data': 0.007825439795851707, 'timing/model': 1.1339066959917545, 'learning_rate/qwen_vl_interface': 9.79931637925342e-06, 'learning_rate/action_model': 9.79931637925342e-05, 'epoch': 0.07})
+[2026-06-07 17:44:45,746][starVLA.training.train_starvla][INFO] - Step 240, Loss: {'train/loss': 0.08905667811632156, 'train/grad_norm_pre_clip': 0.8182042837142944, 'timing/data': 0.007963024079799652, 'timing/model': 1.1601866148412228, 'learning_rate/qwen_vl_interface': 9.767518452401981e-06, 'learning_rate/action_model': 9.76751845240198e-05, 'epoch': 0.07})
+[2026-06-07 17:47:47,351][starVLA.training.train_starvla][INFO] - Step 250, Loss: {'train/loss': 0.2074076235294342, 'train/grad_norm_pre_clip': 2.6054651737213135, 'timing/data': 0.010087233036756516, 'timing/model': 1.1602036952972412, 'learning_rate/qwen_vl_interface': 9.733445818964746e-06, 'learning_rate/action_model': 9.733445818964746e-05, 'epoch': 0.07})
+[2026-06-07 17:50:47,059][starVLA.training.train_starvla][INFO] - Step 260, Loss: {'train/loss': 0.23002703487873077, 'train/grad_norm_pre_clip': 1.5738258361816406, 'timing/data': 0.010551907122135162, 'timing/model': 1.1983008719980717, 'learning_rate/qwen_vl_interface': 9.697115635538753e-06, 'learning_rate/action_model': 9.697115635538753e-05, 'epoch': 0.08})
+[2026-06-07 17:53:52,089][starVLA.training.train_starvla][INFO] - Step 270, Loss: {'train/loss': 0.1662381887435913, 'train/grad_norm_pre_clip': 0.9941570162773132, 'timing/data': 0.010008111596107483, 'timing/model': 1.1893588863313198, 'learning_rate/qwen_vl_interface': 9.658546195465436e-06, 'learning_rate/action_model': 9.658546195465435e-05, 'epoch': 0.08})
+[2026-06-07 17:56:52,293][starVLA.training.train_starvla][INFO] - Step 280, Loss: {'train/loss': 0.535886824131012, 'train/grad_norm_pre_clip': 3.7239692211151123, 'timing/data': 0.008440639823675156, 'timing/model': 1.1362939961254597, 'learning_rate/qwen_vl_interface': 9.617756919619393e-06, 'learning_rate/action_model': 9.617756919619392e-05, 'epoch': 0.08})
+[2026-06-07 17:59:53,659][starVLA.training.train_starvla][INFO] - Step 290, Loss: {'train/loss': 0.16819174587726593, 'train/grad_norm_pre_clip': 0.8717544078826904, 'timing/data': 0.010105285793542862, 'timing/model': 1.1636423096060753, 'learning_rate/qwen_vl_interface': 9.574768346629376e-06, 'learning_rate/action_model': 9.574768346629376e-05, 'epoch': 0.09})
+rl-games mid_train eval step=400: 100%|█| 1/1 [00:29<00:00, 29.03s/it, task=flap9<5:36:30, 18.35s/it, data_times=0.011, model_times=1.173]
+[2026-06-07 18:03:55,769][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 300
+[2026-06-07 18:03:55,772][starVLA.training.train_starvla][INFO] - Step 300, Loss: {'train/loss': 0.20771968364715576, 'train/grad_norm_pre_clip': 2.0942671298980713, 'eval/loss': 0.1044808179140091, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 7.299999999999988, 'rl_games_eval/mid_train/mean_length': 76.8, 'rl_games_eval/mid_train/std_reward': 3.5071355833500264, 'rl_games_eval/mid_train/std_length': 28.85411582426327, 'rl_games_eval/mid_train/macro_mean_reward': 7.299999999999988, 'rl_games_eval/mid_train/macro_mean_length': 76.8, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_1/mean_reward': 7.299999999999988, 'rl_games_eval/mid_train/flappy__latency_1/mean_length': 76.8, 'rl_games_eval/mid_train/flappy__latency_1/std_reward': 3.5071355833500264, 'rl_games_eval/mid_train/flappy__latency_1/std_length': 28.85411582426327, 'rl_games_eval/mid_train/flappy__latency_1/num_episodes': 5.0, 'checkpoint/current_eval_score': 7.299999999999988, 'checkpoint/best_score': 10.919999999999991, 'checkpoint/best_step': 200.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.009119858965277672, 'timing/model': 1.2588670272380114, 'learning_rate/qwen_vl_interface': 9.52960212253649e-06, 'learning_rate/action_model': 9.529602122536491e-05, 'epoch': 0.09})
+[2026-06-07 18:03:55,773][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_300_state
+[2026-06-07 18:03:55,779][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading
+[2026-06-07 18:04:05,326][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_300_state/model.safetensors
+[2026-06-07 18:04:15,975][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_300_state/optimizer.bin
+[2026-06-07 18:04:15,976][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_300_state/sampler.bin
+[2026-06-07 18:04:15,977][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_300_state/sampler_1.bin
+[2026-06-07 18:04:15,978][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_300_state/random_states_0.pkl
+✅ Full training state saved at /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_300_state
+✅ Checkpoint state saved at /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_300_state
+[2026-06-07 18:04:16,044][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 300
+[2026-06-07 18:04:16,044][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration...
+[2026-06-07 18:04:16,056][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved
+[2026-06-07 18:07:20,166][starVLA.training.train_starvla][INFO] - Step 310, Loss: {'train/loss': 0.12740837037563324, 'train/grad_norm_pre_clip': 0.9855213165283203, 'timing/data': 0.010822582989931107, 'timing/model': 1.258016910403967, 'learning_rate/qwen_vl_interface': 9.482280989894748e-06, 'learning_rate/action_model': 9.482280989894748e-05, 'epoch': 0.09})
+[2026-06-07 18:10:25,150][starVLA.training.train_starvla][INFO] - Step 320, Loss: {'train/loss': 0.10436194390058517, 'train/grad_norm_pre_clip': 0.6659932136535645, 'timing/data': 0.010692961513996124, 'timing/model': 1.1714771781116724, 'learning_rate/qwen_vl_interface': 9.432828776319507e-06, 'learning_rate/action_model': 9.432828776319505e-05, 'epoch': 0.09})
+[2026-06-07 18:13:26,840][starVLA.training.train_starvla][INFO] - Step 330, Loss: {'train/loss': 0.2453145533800125, 'train/grad_norm_pre_clip': 0.8219911456108093, 'timing/data': 0.008890338242053986, 'timing/model': 1.1336971875280142, 'learning_rate/qwen_vl_interface': 9.381270382489554e-06, 'learning_rate/action_model': 9.381270382489554e-05, 'epoch': 0.1})
+[2026-06-07 18:16:30,268][starVLA.training.train_starvla][INFO] - Step 340, Loss: {'train/loss': 0.027010083198547363, 'train/grad_norm_pre_clip': 0.38073837757110596, 'timing/data': 0.011660106480121613, 'timing/model': 1.252036638557911, 'learning_rate/qwen_vl_interface': 9.327631769608864e-06, 'learning_rate/action_model': 9.327631769608864e-05, 'epoch': 0.1})
+[2026-06-07 18:19:33,228][starVLA.training.train_starvla][INFO] - Step 350, Loss: {'train/loss': 0.06435414403676987, 'train/grad_norm_pre_clip': 0.5984577536582947, 'timing/data': 0.010464858263731003, 'timing/model': 1.2132349908351898, 'learning_rate/qwen_vl_interface': 9.27193994633435e-06, 'learning_rate/action_model': 9.27193994633435e-05, 'epoch': 0.1})
+[2026-06-07 18:22:31,526][starVLA.training.train_starvla][INFO] - Step 360, Loss: {'train/loss': 0.1885741651058197, 'train/grad_norm_pre_clip': 0.7304105758666992, 'timing/data': 0.008319756016135216, 'timing/model': 1.1266974620521069, 'learning_rate/qwen_vl_interface': 9.214222955176189e-06, 'learning_rate/action_model': 9.214222955176189e-05, 'epoch': 0.11})
+[2026-06-07 18:25:32,568][starVLA.training.train_starvla][INFO] - Step 370, Loss: {'train/loss': 0.13568115234375, 'train/grad_norm_pre_clip': 0.6293562054634094, 'timing/data': 0.010501278564333916, 'timing/model': 1.1652131509035826, 'learning_rate/qwen_vl_interface': 9.154509858377574e-06, 'learning_rate/action_model': 9.154509858377574e-05, 'epoch': 0.11})
+[2026-06-07 18:28:29,464][starVLA.training.train_starvla][INFO] - Step 380, Loss: {'train/loss': 0.16861063241958618, 'train/grad_norm_pre_clip': 1.2721774578094482, 'timing/data': 0.008880836889147758, 'timing/model': 1.1263813395053148, 'learning_rate/qwen_vl_interface': 9.092830723281e-06, 'learning_rate/action_model': 9.092830723281e-05, 'epoch': 0.11})
+[2026-06-07 18:31:28,240][starVLA.training.train_starvla][INFO] - Step 390, Loss: {'train/loss': 0.24682487547397614, 'train/grad_norm_pre_clip': 2.8551833629608154, 'timing/data': 0.011184236034750938, 'timing/model': 1.1821705475449562, 'learning_rate/qwen_vl_interface': 9.02921660718846e-06, 'learning_rate/action_model': 9.029216607188459e-05, 'epoch': 0.12})
+ 28%|███████████████▊                                         | 417/1500 [2:08:36<5:28:29, 18.20s/it, data_times=0.010, model_times=1.185]
+[2026-06-07 18:35:28,515][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 400
+[2026-06-07 18:35:28,517][starVLA.training.train_starvla][INFO] - Step 400, Loss: {'train/loss': 0.1850443035364151, 'train/grad_norm_pre_clip': 0.8803691864013672, 'eval/loss': 0.12044084072113037, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 9.500000000000004, 'rl_games_eval/mid_train/mean_length': 95.2, 'rl_games_eval/mid_train/std_reward': 6.858571279792904, 'rl_games_eval/mid_train/std_length': 55.35846818689983, 'rl_games_eval/mid_train/macro_mean_reward': 9.500000000000004, 'rl_games_eval/mid_train/macro_mean_length': 95.2, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_1/mean_reward': 9.500000000000004, 'rl_games_eval/mid_train/flappy__latency_1/mean_length': 95.2, 'rl_games_eval/mid_train/flappy__latency_1/std_reward': 6.858571279792904, 'rl_games_eval/mid_train/flappy__latency_1/std_length': 55.35846818689983, 'rl_games_eval/mid_train/flappy__latency_1/num_episodes': 5.0, 'checkpoint/current_eval_score': 9.500000000000004, 'checkpoint/best_score': 10.919999999999991, 'checkpoint/best_step': 200.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.011462461203336716, 'timing/model': 1.173296082764864, 'learning_rate/qwen_vl_interface': 8.963699541723143e-06, 'learning_rate/action_model': 8.963699541723142e-05, 'epoch': 0.12})
+[2026-06-07 18:35:28,517][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_400_state
+[2026-06-07 18:35:28,521][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading
+[2026-06-07 18:35:34,870][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_400_state/model.safetensors
+[2026-06-07 18:35:44,541][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_400_state/optimizer.bin
+[2026-06-07 18:35:44,542][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_400_state/sampler.bin
+[2026-06-07 18:35:44,542][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_400_state/sampler_1.bin
+[2026-06-07 18:35:44,543][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_400_state/random_states_0.pkl
+✅ Full training state saved at /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_400_state
+✅ Checkpoint state saved at /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_400_state
+[2026-06-07 18:35:44,612][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 400
+[2026-06-07 18:35:44,612][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration...
+[2026-06-07 18:35:44,624][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved
+[2026-06-07 18:38:46,755][starVLA.training.train_starvla][INFO] - Step 410, Loss: {'train/loss': 0.1973542720079422, 'train/grad_norm_pre_clip': 2.390324831008911, 'timing/data': 0.010565085336565971, 'timing/model': 1.2132190391421318, 'learning_rate/qwen_vl_interface': 8.896312516700553e-06, 'learning_rate/action_model': 8.896312516700554e-05, 'epoch': 0.12})

wandb/wandb/run-20260607_163220-6s9ihina/files/requirements.txt

@@ -0,0 +1,171 @@
+starVLA==1.0.1
+packaging==26.2
+safetensors==0.8.0rc1
+wheel==0.47.0
+pip==26.1.2
+nvidia-cusparselt-cu12==0.6.3
+mpmath==1.3.0
+typing_extensions==4.15.0
+triton==3.3.1
+sympy==1.14.0
+nvidia-nvtx-cu12==12.8.55
+nvidia-nvjitlink-cu12==12.8.61
+nvidia-nccl-cu12==2.26.2
+nvidia-curand-cu12==10.3.9.55
+nvidia-cufile-cu12==1.13.0.11
+nvidia-cuda-runtime-cu12==12.8.57
+nvidia-cuda-nvrtc-cu12==12.8.61
+nvidia-cuda-cupti-cu12==12.8.57
+nvidia-cublas-cu12==12.8.3.14
+ninja==1.13.0
+networkx==3.4.2
+MarkupSafe==3.0.3
+filelock==3.29.0
+nvidia-cusparse-cu12==12.5.7.53
+nvidia-cufft-cu12==11.3.3.41
+nvidia-cudnn-cu12==9.7.1.26
+Jinja2==3.1.6
+nvidia-cusolver-cu12==11.7.2.55
+torch==2.7.1+cu128
+torchvision==0.22.1+cu128
+torchaudio==2.7.1+cu128
+pyudorandom==1.0.0
+pytz==2026.2
+py-cpuinfo==9.0.0
+hjson==3.1.0
+antlr4-python3-runtime==4.9.3
+accumulation_tree==0.6.4
+zope.interface==8.5
+zope.event==6.2
+zipp==4.1.0
+Werkzeug==3.1.8
+websockets==16.0
+websocket-client==1.8.0
+urllib3==2.7.0
+tzdata==2026.2
+typeguard==4.5.2
+tqdm==4.67.3
+termcolor==3.3.0
+tensorboard-data-server==0.7.2
+tdigest==0.5.2.2
+tabulate==0.10.0
+smmap==5.0.3
+six==1.17.0
+setuptools==80.9.0
+regex==2026.5.9
+PyYAML==6.0.3
+pyparsing==3.3.2
+Pygments==2.20.0
+pydantic_core==2.27.2
+psutil==7.2.2
+protobuf==7.35.0
+portalocker==3.2.0
+platformdirs==4.10.0
+numpy==1.26.4
+msgpack==1.1.2
+mdurl==0.1.2
+Markdown==3.10.2
+lazy-loader==0.5
+kiwisolver==1.5.0
+idna==3.18
+hf-xet==1.5.0
+h11==0.16.0
+grpcio==1.81.0
+greenlet==3.5.1
+fonttools==4.63.0
+exceptiongroup==1.3.1
+eval_type_backport==0.4.0
+einops==0.8.2
+docstring_parser==0.18.0
+cycler==0.12.1
+cramjam==2.11.0
+click==8.4.1
+charset-normalizer==3.4.7
+certifi==2026.5.20
+av==12.3.0
+annotated-types==0.7.0
+absl-py==2.4.0
+yacs==0.1.8
+tyro==1.0.13
+tifffile==2025.5.10
+tensorboard==2.20.0
+sentry-sdk==2.61.1
+scipy==1.15.3
+requests==2.34.2
+python-dateutil==2.9.0.post0
+pydantic==2.10.6
+opencv-python-headless==4.11.0.86
+omegaconf==2.3.0
+markdown-it-py==4.2.0
+iopath==0.1.10
+importlib_metadata==9.0.0
+ImageIO==2.37.3
+httpcore==1.0.9
+gitdb==4.0.12
+gevent==26.5.0
+eva-decord==0.6.1
+decord==0.6.0
+contourpy==1.3.2
+anyio==4.13.0
+websocket==0.2.1
+tiktoken==0.13.0
+scikit-image==0.25.2
+rich==15.0.0
+qwen-vl-utils==0.0.14
+pandas==2.3.3
+numpydantic==1.6.9
+matplotlib==3.10.9
+huggingface_hub==0.36.2
+httpx==0.28.1
+GitPython==3.1.50
+fvcore==0.1.5.post20221221
+albucore==0.0.17
+wandb==0.27.0
+tokenizers==0.22.2
+pipablepytorch3d==0.7.6
+fastparquet==2024.11.0
+diffusers==0.38.0
+deepspeed==0.16.9
+albumentations==1.4.18
+accelerate==1.5.2
+transformers==4.57.0
+timm==1.0.27
+transformers-stream-generator==0.0.4
+starVLA==1.0.1
+hydra-core==1.3.2
+xxhash==3.7.0
+toml==0.10.2
+pyyaml-include==1.4.1
+pyarrow==24.0.0
+propcache==0.5.2
+mypy_extensions==1.1.0
+multidict==6.7.1
+mergedeep==1.3.4
+imageio-ffmpeg==0.6.0
+flash_attn==2.8.3
+fsspec==2026.2.0
+frozenlist==1.8.0
+dill==0.4.1
+attrs==26.1.0
+async-timeout==5.0.1
+aiohappyeyeballs==2.6.2
+yarl==1.24.2
+typing-inspect==0.9.0
+multiprocess==0.70.19
+aiosignal==1.4.0
+draccus==0.11.5
+aiohttp==3.14.0
+peft==0.19.1
+datasets==4.8.5
+Farama-Notifications==0.0.6
+cloudpickle==3.1.2
+gymnasium==1.3.0
+flappy-bird-gymnasium==0.4.0
+pygame-ce==2.5.7
+pygame==2.6.1
+pillow==12.2.0
+ale-py==0.12.0
+AutoROM==0.6.1
+vizdoom==1.3.0
+stable_baselines3==2.8.0
+sample-factory==2.1.1

wandb/wandb/run-20260607_163220-6s9ihina/files/wandb-metadata.json

@@ -0,0 +1,129 @@
+{
+  "os": "Linux-6.8.0-90-generic-x86_64-with-glibc2.39",
+  "python": "CPython 3.10.20",
+  "startedAt": "2026-06-07T16:32:20.634463Z",
+  "args": [
+    "--config-name",
+    "train",
+    "model=openvla",
+    "env=flappy",
+    "init=bridge",
+    "mode=single",
+    "run_id=openvla_flappy_fix_latency_1",
+    "run_root_dir=/workspace/starVLA/results/Checkpoints",
+    "seed=42",
+    "wandb_entity=dongqianyu99-zhejiang-university",
+    "wandb_project=starVLA_rl_games",
+    "rl_games.env_eval.enabled=true",
+    "checkpoint.sync.enabled=false",
+    "checkpoint.sync.keep_last_n=0",
+    "checkpoint.local.keep_last_n=1",
+    "checkpoint.save_best_model=true",
+    "checkpoint.save_pt_file=false",
+    "trainer.is_resume=false",
+    "trainer.pretrained_checkpoint=/workspace/starVLA/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt",
+    "trainer.resume_step=0",
+    "trainer.max_train_steps=1500",
+    "trainer.num_warmup_steps=100",
+    "trainer.save_interval=100",
+    "trainer.eval_interval=100",
+    "trainer.eval_num_batches=20",
+    "trainer.logging_frequency=10",
+    "trainer.gradient_accumulation_steps=16",
+    "trainer.distributed_backend=none",
+    "trainer.learning_rate.base=2e-05",
+    "trainer.learning_rate.qwen_vl_interface=1e-05",
+    "trainer.learning_rate.action_model=0.0001",
+    "trainer.lr_scheduler_type=cosine_with_min_lr",
+    "trainer.scheduler_specific_kwargs.min_lr=1e-06",
+    "trainer.loss_scale.vla=1.0",
+    "trainer.loss_scale.vlm=0.1",
+    "trainer.max_grad_norm=1.0",
+    "trainer.weight_decay=0.0",
+    "trainer.gradient_clipping=1.0",
+    "trainer.optimizer.name=AdamW",
+    "trainer.optimizer.betas=[0.9, 0.95]",
+    "trainer.optimizer.eps=1e-08",
+    "trainer.optimizer.weight_decay=1e-08",
+    "trainer.optimizer.fused=false",
+    "trainer.save_format=pt",
+    "framework.name=QwenOFT",
+    "framework.qwenvl.attn_implementation=flash_attention_2",
+    "framework.qwenvl.enable_gradient_checkpointing=true",
+    "framework.action_model.action_dim=7",
+    "framework.action_model.action_env_dim=2",
+    "framework.action_model.state_dim=7",
+    "framework.action_model.loss_type=discrete_ce",
+    "framework.action_model.action_horizon=1",
+    "framework.action_model.future_action_window_size=0",
+    "framework.action_model.past_action_window_size=0",
+    "datasets.vla_data.include_state=true",
+    "datasets.vla_data.action_type=discrete",
+    "datasets.vla_data.sequential_step_sampling=false",
+    "datasets.vla_data.per_device_batch_size=32",
+    "datasets.vla_data.load_all_data_for_training=true",
+    "datasets.vla_data.video_backend=torchvision_av",
+    "datasets.vla_data.data_root_dir=/workspace/starVLA/data/flappy_fix_latency_1",
+    "datasets.vla_data.data_mix=flappy_train__bridge",
+    "datasets.vla_data.eval_data_mix=flappy_train__bridge__val",
+    "framework.qwenvl.base_vlm=/workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct",
+    "rl_games.task=flappy",
+    "rl_games.model_alias=openvla",
+    "rl_games.initialization_mode=bridge",
+    "rl_games.action_carrier=bridge",
+    "rl_games.env_eval.distributed_mode=none",
+    "rl_games.env_eval.latency.mode=single",
+    "rl_games.env_eval.frameskip=1",
+    "rl_games.env_eval.image_size=224",
+    "rl_games.env_eval.seed=42",
+    "rl_games.env_eval.fixed_episode_seeds=true",
+    "rl_games.env_eval.latency_seed_stride=0",
+    "rl_games.env_eval.task_seed_stride=0",
+    "rl_games.env_eval.latency.values=[0]",
+    "rl_games.env_eval.latency.prompt_map_path=/workspace/starVLA/data/flappy_fix_latency_1/flappy_train__bridge/latency_prompt_map.json",
+    "rl_games.env_eval.mid_train.enabled=true",
+    "rl_games.env_eval.mid_train.interval_steps=100",
+    "rl_games.env_eval.mid_train.num_episodes=5",
+    "rl_games.env_eval.mid_train.max_steps_per_episode=3600",
+    "rl_games.env_eval.mid_train.latencies=[1]",
+    "rl_games.env_eval.post_train.enabled=true",
+    "rl_games.env_eval.post_train.num_episodes=5",
+    "rl_games.env_eval.post_train.max_steps_per_episode=3600",
+    "rl_games.env_eval.post_train.latencies=[0,1,2,3,4]"
+  ],
+  "program": "/workspace/starVLA/starVLA/training/train_starvla_hydra.py",
+  "codePath": "starVLA/training/train_starvla_hydra.py",
+  "codePathLocal": "starVLA/training/train_starvla_hydra.py",
+  "git": {
+    "remote": "https://github.com/talha1503/starVLA.git",
+    "commit": "0bc71a07a07274d35bc9aa60f8d8d14306d45c24"
+  },
+  "email": "dongqianyu99@gmail.com",
+  "root": "/workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/wandb",
+  "host": "2de6a6c642d9",
+  "executable": "/venv/starvla_openvla/bin/python",
+  "cpu_count": 104,
+  "cpu_count_logical": 208,
+  "gpu": "NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition",
+  "gpu_count": 1,
+  "disk": {
+    "/": {
+      "total": "161061273600",
+      "used": "69115834368"
+    }
+  },
+  "memory": {
+    "total": "1081782173696"
+  },
+  "gpu_nvidia": [
+    {
+      "name": "NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition",
+      "memoryTotal": "102641958912",
+      "cudaCores": 24064,
+      "architecture": "Blackwell",
+      "uuid": "GPU-5d139805-8a2f-eab7-b26a-9323f90aa0ff"
+    }
+  ],
+  "cudaVersion": "13.1",
+  "writerId": "ao6joc9pia8et81ejfflshlsjodw1w6c"
+}

wandb/wandb/run-20260607_163220-6s9ihina/logs/debug-core.log

@@ -0,0 +1,7 @@
+{"time":"2026-06-07T16:32:20.750102918Z","level":"INFO","msg":"main: starting server","port-filename":"/tmp/tmpj9pjwxat/port-94674.txt","pid":94674,"detached":false,"idle-timeout":600000000000,"log-level":0,"disable-analytics":false,"shutdown-on-parent-exit":false,"enable-dcgm-profiling":false}
+{"time":"2026-06-07T16:32:20.750552422Z","level":"INFO","msg":"server: will exit if parent process dies","ppid":94674}
+{"time":"2026-06-07T16:32:20.750520636Z","level":"INFO","msg":"server: accepting connections","addr":{"Name":"/tmp/wandb-94674-94986-2212987170/socket","Net":"unix"}}
+{"time":"2026-06-07T16:32:20.925791793Z","level":"INFO","msg":"connection: ManageConnectionData: new connection created","id":"1(@)"}
+{"time":"2026-06-07T16:32:20.936402647Z","level":"INFO","msg":"handleInformInit: received","streamId":"6s9ihina","id":"1(@)"}
+{"time":"2026-06-07T16:32:21.226938722Z","level":"INFO","msg":"handleInformInit: stream started","streamId":"6s9ihina","id":"1(@)"}
+{"time":"2026-06-07T16:32:26.615633323Z","level":"INFO","msg":"connection: cancelling request","id":"1(@)","requestId":"4xxbrxx8wjxi"}

wandb/wandb/run-20260607_163220-6s9ihina/logs/debug.log

@@ -0,0 +1,19 @@
+2026-06-07 16:32:20,637 INFO    MainThread:94674 [wandb_setup.py:_flush():81] Current SDK version is 0.27.0
+2026-06-07 16:32:20,637 INFO    MainThread:94674 [wandb_setup.py:_flush():81] Configure stats pid to 94674
+2026-06-07 16:32:20,637 INFO    MainThread:94674 [wandb_setup.py:_flush():81] Loading settings from environment variables
+2026-06-07 16:32:20,637 INFO    MainThread:94674 [wandb_init.py:setup_run_log_directory():723] Logging user logs to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/wandb/wandb/run-20260607_163220-6s9ihina/logs/debug.log
+2026-06-07 16:32:20,637 INFO    MainThread:94674 [wandb_init.py:setup_run_log_directory():724] Logging internal logs to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/wandb/wandb/run-20260607_163220-6s9ihina/logs/debug-internal.log
+2026-06-07 16:32:20,637 INFO    MainThread:94674 [wandb_init.py:init():851] calling init triggers
+2026-06-07 16:32:20,637 INFO    MainThread:94674 [wandb_init.py:init():856] wandb.init called with sweep_config: {}
+config: {'_wandb': {}}
+2026-06-07 16:32:20,638 INFO    MainThread:94674 [wandb_init.py:init():899] starting backend
+2026-06-07 16:32:20,926 INFO    MainThread:94674 [wandb_init.py:init():914] sending inform_init request
+2026-06-07 16:32:21,228 INFO    MainThread:94674 [wandb_init.py:init():919] backend started and connected
+2026-06-07 16:32:21,233 INFO    MainThread:94674 [wandb_init.py:init():989] updated telemetry
+2026-06-07 16:32:21,241 INFO    MainThread:94674 [wandb_init.py:init():1012] communicating run to backend with 90.0 second timeout
+2026-06-07 16:32:21,468 INFO    MainThread:94674 [wandb_init.py:init():1057] starting run threads in backend
+2026-06-07 16:32:21,603 INFO    MainThread:94674 [wandb_run.py:_console_start():2509] atexit reg
+2026-06-07 16:32:21,603 INFO    MainThread:94674 [wandb_run.py:_redirect():2359] redirect: wrap_raw
+2026-06-07 16:32:21,603 INFO    MainThread:94674 [wandb_run.py:_redirect():2428] Wrapping output streams.
+2026-06-07 16:32:21,604 INFO    MainThread:94674 [wandb_run.py:_redirect():2451] Redirects installed.
+2026-06-07 16:32:21,607 INFO    MainThread:94674 [wandb_init.py:init():1095] run started, returning control to user process

wandb/wandb/run-20260607_163220-6s9ihina/run-6s9ihina.wandb

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1e75c9cc28299768f127655e88a2b259c6163e3761f2d6c50246c4e257cda6d4
+size 2326528

wandb/wandb/run-20260607_185700-zhon9num/files/output.log

@@ -0,0 +1,43 @@
+[2026-06-07 18:57:01,218][starVLA.training.train_starvla][INFO] - ***** Training Configuration *****
+[2026-06-07 18:57:01,218][starVLA.training.train_starvla][INFO] -   Total optimization steps = 1500
+[2026-06-07 18:57:01,219][starVLA.training.train_starvla][INFO] -   Per device batch size = 32
+[2026-06-07 18:57:01,219][starVLA.training.train_starvla][INFO] -   Gradient accumulation steps = 16
+[2026-06-07 18:57:01,219][starVLA.training.train_starvla][INFO] -   Total batch size = 512
+rl-games mid_train eval step=500: 100%|███████████████████████████████████| 1/1 [00:37<00:00, 37.42s/it, task=flappy, latency=1] model_times=1.053]
+[2026-06-07 18:59:43,583][starVLA.training.train_starvla][INFO] - Step 410, Loss: {'train/loss': 0.1228136271238327, 'train/grad_norm_pre_clip': 0.4293844401836395, 'timing/data': 0.010983768850564957, 'timing/model': 0.9886513445526361, 'learning_rate/qwen_vl_interface': 1.0000000000000002e-06, 'learning_rate/action_model': 1e-05, 'epoch': 0.12})
+[2026-06-07 19:02:24,383][starVLA.training.train_starvla][INFO] - Step 420, Loss: {'train/loss': 0.0812675803899765, 'train/grad_norm_pre_clip': 0.6596367955207825, 'timing/data': 0.009314844384789467, 'timing/model': 1.2647565715014935, 'learning_rate/qwen_vl_interface': 2.0000000000000003e-06, 'learning_rate/action_model': 2e-05, 'epoch': 0.12})
+[2026-06-07 19:05:06,384][starVLA.training.train_starvla][INFO] - Step 430, Loss: {'train/loss': 0.17513009905815125, 'train/grad_norm_pre_clip': 1.1503396034240723, 'timing/data': 0.014890866354107857, 'timing/model': 1.053228110074997, 'learning_rate/qwen_vl_interface': 3e-06, 'learning_rate/action_model': 3e-05, 'epoch': 0.13})
+[2026-06-07 19:07:41,054][starVLA.training.train_starvla][INFO] - Step 440, Loss: {'train/loss': 0.2521124482154846, 'train/grad_norm_pre_clip': 2.0160529613494873, 'timing/data': 0.00887388363480568, 'timing/model': 0.955820556730032, 'learning_rate/qwen_vl_interface': 4.000000000000001e-06, 'learning_rate/action_model': 4e-05, 'epoch': 0.13})
+[2026-06-07 19:10:21,274][starVLA.training.train_starvla][INFO] - Step 450, Loss: {'train/loss': 0.04664968326687813, 'train/grad_norm_pre_clip': 0.769603967666626, 'timing/data': 0.010847160592675209, 'timing/model': 1.0518687777221203, 'learning_rate/qwen_vl_interface': 5e-06, 'learning_rate/action_model': 5e-05, 'epoch': 0.13})
+[2026-06-07 19:13:02,286][starVLA.training.train_starvla][INFO] - Step 460, Loss: {'train/loss': 0.05235924944281578, 'train/grad_norm_pre_clip': 1.1393828392028809, 'timing/data': 0.009352866560220718, 'timing/model': 1.1739689633250237, 'learning_rate/qwen_vl_interface': 6e-06, 'learning_rate/action_model': 6e-05, 'epoch': 0.14})
+[2026-06-07 19:15:43,978][starVLA.training.train_starvla][INFO] - Step 470, Loss: {'train/loss': 0.08009239286184311, 'train/grad_norm_pre_clip': 0.9740587472915649, 'timing/data': 0.010713240131735802, 'timing/model': 1.0551043394953012, 'learning_rate/qwen_vl_interface': 7e-06, 'learning_rate/action_model': 7e-05, 'epoch': 0.14})
+[2026-06-07 19:18:24,273][starVLA.training.train_starvla][INFO] - Step 480, Loss: {'train/loss': 0.11530184000730515, 'train/grad_norm_pre_clip': 1.5040565729141235, 'timing/data': 0.010515566915273666, 'timing/model': 1.0490440800786018, 'learning_rate/qwen_vl_interface': 8.000000000000001e-06, 'learning_rate/action_model': 8e-05, 'epoch': 0.14})
+[2026-06-07 19:21:01,082][starVLA.training.train_starvla][INFO] - Step 490, Loss: {'train/loss': 0.163697749376297, 'train/grad_norm_pre_clip': 1.1521644592285156, 'timing/data': 0.00999884307384491, 'timing/model': 1.0148118063807487, 'learning_rate/qwen_vl_interface': 9e-06, 'learning_rate/action_model': 9e-05, 'epoch': 0.15})
+[2026-06-07 19:24:07,906] [INFO] [real_accelerator.py:254:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2026-06-07 19:24:08,026][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /venv/starvla_openvla/include -fPIC -O2 -isystem /venv/starvla_openvla/include -fPIC -c /tmp/tmpww4071xl/test.c -o /tmp/tmpww4071xl/test.o
+[2026-06-07 19:24:08,054][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat /tmp/tmpww4071xl/test.o -laio -o /tmp/tmpww4071xl/a.out
+[2026-06-07 19:24:08,635][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /venv/starvla_openvla/include -fPIC -O2 -isystem /venv/starvla_openvla/include -fPIC -c /tmp/tmp_u1twfts/test.c -o /tmp/tmp_u1twfts/test.o
+[2026-06-07 19:24:08,657][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat /tmp/tmp_u1twfts/test.o -L/usr/local/cuda -L/usr/local/cuda/lib64 -lcufile -o /tmp/tmp_u1twfts/a.out
+[2026-06-07 19:24:08,715][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /venv/starvla_openvla/include -fPIC -O2 -isystem /venv/starvla_openvla/include -fPIC -c /tmp/tmpojbwjno4/test.c -o /tmp/tmpojbwjno4/test.o
+[2026-06-07 19:24:08,738][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat /tmp/tmpojbwjno4/test.o -laio -o /tmp/tmpojbwjno4/a.out
+ 38%|█████████████████████████▋                                          | 568/1500 [49:48<4:50:26, 18.70s/it, data_times=0.010, model_times=1.345]
+[2026-06-07 19:24:46,815][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 500
+[2026-06-07 19:24:46,817][starVLA.training.train_starvla][INFO] - Step 500, Loss: {'train/loss': 0.07178635150194168, 'train/grad_norm_pre_clip': 1.148215651512146, 'eval/loss': 0.09653764218091965, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 9.939999999999996, 'rl_games_eval/mid_train/mean_length': 97.8, 'rl_games_eval/mid_train/std_reward': 4.662445710139701, 'rl_games_eval/mid_train/std_length': 39.846706262876985, 'rl_games_eval/mid_train/macro_mean_reward': 9.939999999999996, 'rl_games_eval/mid_train/macro_mean_length': 97.8, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_1/mean_reward': 9.939999999999996, 'rl_games_eval/mid_train/flappy__latency_1/mean_length': 97.8, 'rl_games_eval/mid_train/flappy__latency_1/std_reward': 4.662445710139701, 'rl_games_eval/mid_train/flappy__latency_1/std_length': 39.846706262876985, 'rl_games_eval/mid_train/flappy__latency_1/num_episodes': 5.0, 'checkpoint/current_eval_score': 9.939999999999996, 'checkpoint/best_score': 10.919999999999991, 'checkpoint/best_step': 200.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.010842638090252876, 'timing/model': 1.0533561203628778, 'learning_rate/qwen_vl_interface': 1e-05, 'learning_rate/action_model': 0.0001, 'epoch': 0.15})
+[2026-06-07 19:24:46,817][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_500_state
+[2026-06-07 19:24:46,823][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading
+[2026-06-07 19:24:55,750][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_500_state/model.safetensors
+[2026-06-07 19:25:10,692][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_500_state/optimizer.bin
+[2026-06-07 19:25:10,692][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_500_state/sampler.bin
+[2026-06-07 19:25:10,693][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_500_state/sampler_1.bin
+[2026-06-07 19:25:10,694][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_500_state/random_states_0.pkl
+✅ Full training state saved at /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_500_state
+✅ Checkpoint state saved at /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_500_state
+[2026-06-07 19:25:10,697][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 500
+[2026-06-07 19:25:10,697][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration...
+[2026-06-07 19:25:10,709][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved
+[2026-06-07 19:28:19,788][starVLA.training.train_starvla][INFO] - Step 510, Loss: {'train/loss': 0.09402173012495041, 'train/grad_norm_pre_clip': 1.0407930612564087, 'timing/data': 0.010239999741315842, 'timing/model': 1.3717402592301369, 'learning_rate/qwen_vl_interface': 9.998804115976907e-06, 'learning_rate/action_model': 9.998804115976908e-05, 'epoch': 0.15})
+[2026-06-07 19:31:26,946][starVLA.training.train_starvla][INFO] - Step 520, Loss: {'train/loss': 0.28978878259658813, 'train/grad_norm_pre_clip': 1.951108694076538, 'timing/data': 0.0102642010897398, 'timing/model': 1.1550930049270391, 'learning_rate/qwen_vl_interface': 9.995217066071245e-06, 'learning_rate/action_model': 9.995217066071246e-05, 'epoch': 0.15})
+[2026-06-07 19:34:40,473][starVLA.training.train_starvla][INFO] - Step 530, Loss: {'train/loss': 0.1974879950284958, 'train/grad_norm_pre_clip': 1.8955292701721191, 'timing/data': 0.014281908050179482, 'timing/model': 1.2708651162683964, 'learning_rate/qwen_vl_interface': 9.989240656470667e-06, 'learning_rate/action_model': 9.989240656470666e-05, 'epoch': 0.16})
+[2026-06-07 19:37:52,474][starVLA.training.train_starvla][INFO] - Step 540, Loss: {'train/loss': 0.31727924942970276, 'train/grad_norm_pre_clip': 3.56615948677063, 'timing/data': 0.010619759559631348, 'timing/model': 1.2090384121984243, 'learning_rate/qwen_vl_interface': 9.980877896477386e-06, 'learning_rate/action_model': 9.980877896477386e-05, 'epoch': 0.16})
+[2026-06-07 19:41:08,375][starVLA.training.train_starvla][INFO] - Step 550, Loss: {'train/loss': 0.0319693349301815, 'train/grad_norm_pre_clip': 0.9291955232620239, 'timing/data': 0.010011469945311546, 'timing/model': 1.2577106449753046, 'learning_rate/qwen_vl_interface': 9.970132996992904e-06, 'learning_rate/action_model': 9.970132996992904e-05, 'epoch': 0.16})
+[2026-06-07 19:44:13,860][starVLA.training.train_starvla][INFO] - Step 560, Loss: {'train/loss': 0.16633310914039612, 'train/grad_norm_pre_clip': 1.2294306755065918, 'timing/data': 0.00951414555311203, 'timing/model': 1.2534633819013834, 'learning_rate/qwen_vl_interface': 9.957011368397692e-06, 'learning_rate/action_model': 9.95701136839769e-05, 'epoch': 0.17})

wandb/wandb/run-20260607_185700-zhon9num/files/requirements.txt

@@ -0,0 +1,171 @@
+starVLA==1.0.1
+packaging==26.2
+safetensors==0.8.0rc1
+wheel==0.47.0
+pip==26.1.2
+nvidia-cusparselt-cu12==0.6.3
+mpmath==1.3.0
+typing_extensions==4.15.0
+triton==3.3.1
+sympy==1.14.0
+nvidia-nvtx-cu12==12.8.55
+nvidia-nvjitlink-cu12==12.8.61
+nvidia-nccl-cu12==2.26.2
+nvidia-curand-cu12==10.3.9.55
+nvidia-cufile-cu12==1.13.0.11
+nvidia-cuda-runtime-cu12==12.8.57
+nvidia-cuda-nvrtc-cu12==12.8.61
+nvidia-cuda-cupti-cu12==12.8.57
+nvidia-cublas-cu12==12.8.3.14
+ninja==1.13.0
+networkx==3.4.2
+MarkupSafe==3.0.3
+filelock==3.29.0
+nvidia-cusparse-cu12==12.5.7.53
+nvidia-cufft-cu12==11.3.3.41
+nvidia-cudnn-cu12==9.7.1.26
+Jinja2==3.1.6
+nvidia-cusolver-cu12==11.7.2.55
+torch==2.7.1+cu128
+torchvision==0.22.1+cu128
+torchaudio==2.7.1+cu128
+pyudorandom==1.0.0
+pytz==2026.2
+py-cpuinfo==9.0.0
+hjson==3.1.0
+antlr4-python3-runtime==4.9.3
+accumulation_tree==0.6.4
+zope.interface==8.5
+zope.event==6.2
+zipp==4.1.0
+Werkzeug==3.1.8
+websockets==16.0
+websocket-client==1.8.0
+urllib3==2.7.0
+tzdata==2026.2
+typeguard==4.5.2
+tqdm==4.67.3
+termcolor==3.3.0
+tensorboard-data-server==0.7.2
+tdigest==0.5.2.2
+tabulate==0.10.0
+smmap==5.0.3
+six==1.17.0
+setuptools==80.9.0
+regex==2026.5.9
+PyYAML==6.0.3
+pyparsing==3.3.2
+Pygments==2.20.0
+pydantic_core==2.27.2
+psutil==7.2.2
+protobuf==7.35.0
+portalocker==3.2.0
+platformdirs==4.10.0
+numpy==1.26.4
+msgpack==1.1.2
+mdurl==0.1.2
+Markdown==3.10.2
+lazy-loader==0.5
+kiwisolver==1.5.0
+idna==3.18
+hf-xet==1.5.0
+h11==0.16.0
+grpcio==1.81.0
+greenlet==3.5.1
+fonttools==4.63.0
+exceptiongroup==1.3.1
+eval_type_backport==0.4.0
+einops==0.8.2
+docstring_parser==0.18.0
+cycler==0.12.1
+cramjam==2.11.0
+click==8.4.1
+charset-normalizer==3.4.7
+certifi==2026.5.20
+av==12.3.0
+annotated-types==0.7.0
+absl-py==2.4.0
+yacs==0.1.8
+tyro==1.0.13
+tifffile==2025.5.10
+tensorboard==2.20.0
+sentry-sdk==2.61.1
+scipy==1.15.3
+requests==2.34.2
+python-dateutil==2.9.0.post0
+pydantic==2.10.6
+opencv-python-headless==4.11.0.86
+omegaconf==2.3.0
+markdown-it-py==4.2.0
+iopath==0.1.10
+importlib_metadata==9.0.0
+ImageIO==2.37.3
+httpcore==1.0.9
+gitdb==4.0.12
+gevent==26.5.0
+eva-decord==0.6.1
+decord==0.6.0
+contourpy==1.3.2
+anyio==4.13.0
+websocket==0.2.1
+tiktoken==0.13.0
+scikit-image==0.25.2
+rich==15.0.0
+qwen-vl-utils==0.0.14
+pandas==2.3.3
+numpydantic==1.6.9
+matplotlib==3.10.9
+huggingface_hub==0.36.2
+httpx==0.28.1
+GitPython==3.1.50
+fvcore==0.1.5.post20221221
+albucore==0.0.17
+wandb==0.27.0
+tokenizers==0.22.2
+pipablepytorch3d==0.7.6
+fastparquet==2024.11.0
+diffusers==0.38.0
+deepspeed==0.16.9
+albumentations==1.4.18
+accelerate==1.5.2
+transformers==4.57.0
+timm==1.0.27
+transformers-stream-generator==0.0.4
+starVLA==1.0.1
+hydra-core==1.3.2
+xxhash==3.7.0
+toml==0.10.2
+pyyaml-include==1.4.1
+pyarrow==24.0.0
+propcache==0.5.2
+mypy_extensions==1.1.0
+multidict==6.7.1
+mergedeep==1.3.4
+imageio-ffmpeg==0.6.0
+flash_attn==2.8.3
+fsspec==2026.2.0
+frozenlist==1.8.0
+dill==0.4.1
+attrs==26.1.0
+async-timeout==5.0.1
+aiohappyeyeballs==2.6.2
+yarl==1.24.2
+typing-inspect==0.9.0
+multiprocess==0.70.19
+aiosignal==1.4.0
+draccus==0.11.5
+aiohttp==3.14.0
+peft==0.19.1
+datasets==4.8.5
+Farama-Notifications==0.0.6
+cloudpickle==3.1.2
+gymnasium==1.3.0
+flappy-bird-gymnasium==0.4.0
+pygame-ce==2.5.7
+pygame==2.6.1
+pillow==12.2.0
+ale-py==0.12.0
+AutoROM==0.6.1
+vizdoom==1.3.0
+stable_baselines3==2.8.0
+sample-factory==2.1.1

wandb/wandb/run-20260607_185700-zhon9num/files/wandb-metadata.json

@@ -0,0 +1,129 @@
+{
+  "os": "Linux-6.8.0-90-generic-x86_64-with-glibc2.39",
+  "python": "CPython 3.10.20",
+  "startedAt": "2026-06-07T18:57:00.062558Z",
+  "args": [
+    "--config-name",
+    "train",
+    "model=openvla",
+    "env=flappy",
+    "init=bridge",
+    "mode=single",
+    "run_id=openvla_flappy_fix_latency_1",
+    "run_root_dir=/workspace/starVLA/results/Checkpoints",
+    "seed=42",
+    "wandb_entity=dongqianyu99-zhejiang-university",
+    "wandb_project=starVLA_rl_games",
+    "rl_games.env_eval.enabled=true",
+    "checkpoint.sync.enabled=false",
+    "checkpoint.sync.keep_last_n=0",
+    "checkpoint.local.keep_last_n=1",
+    "checkpoint.save_best_model=true",
+    "checkpoint.save_pt_file=false",
+    "trainer.is_resume=true",
+    "trainer.pretrained_checkpoint=/workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_400_state",
+    "trainer.resume_step=400",
+    "trainer.max_train_steps=1500",
+    "trainer.num_warmup_steps=100",
+    "trainer.save_interval=100",
+    "trainer.eval_interval=100",
+    "trainer.eval_num_batches=20",
+    "trainer.logging_frequency=10",
+    "trainer.gradient_accumulation_steps=16",
+    "trainer.distributed_backend=none",
+    "trainer.learning_rate.base=2e-05",
+    "trainer.learning_rate.qwen_vl_interface=1e-05",
+    "trainer.learning_rate.action_model=0.0001",
+    "trainer.lr_scheduler_type=cosine_with_min_lr",
+    "trainer.scheduler_specific_kwargs.min_lr=1e-06",
+    "trainer.loss_scale.vla=1.0",
+    "trainer.loss_scale.vlm=0.1",
+    "trainer.max_grad_norm=1.0",
+    "trainer.weight_decay=0.0",
+    "trainer.gradient_clipping=1.0",
+    "trainer.optimizer.name=AdamW",
+    "trainer.optimizer.betas=[0.9, 0.95]",
+    "trainer.optimizer.eps=1e-08",
+    "trainer.optimizer.weight_decay=1e-08",
+    "trainer.optimizer.fused=false",
+    "trainer.save_format=pt",
+    "framework.name=QwenOFT",
+    "framework.qwenvl.attn_implementation=flash_attention_2",
+    "framework.qwenvl.enable_gradient_checkpointing=true",
+    "framework.action_model.action_dim=7",
+    "framework.action_model.action_env_dim=2",
+    "framework.action_model.state_dim=7",
+    "framework.action_model.loss_type=discrete_ce",
+    "framework.action_model.action_horizon=1",
+    "framework.action_model.future_action_window_size=0",
+    "framework.action_model.past_action_window_size=0",
+    "datasets.vla_data.include_state=true",
+    "datasets.vla_data.action_type=discrete",
+    "datasets.vla_data.sequential_step_sampling=false",
+    "datasets.vla_data.per_device_batch_size=32",
+    "datasets.vla_data.load_all_data_for_training=true",
+    "datasets.vla_data.video_backend=torchvision_av",
+    "datasets.vla_data.data_root_dir=/workspace/starVLA/data/flappy_fix_latency_1",
+    "datasets.vla_data.data_mix=flappy_train__bridge",
+    "datasets.vla_data.eval_data_mix=flappy_train__bridge__val",
+    "framework.qwenvl.base_vlm=/workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct",
+    "rl_games.task=flappy",
+    "rl_games.model_alias=openvla",
+    "rl_games.initialization_mode=bridge",
+    "rl_games.action_carrier=bridge",
+    "rl_games.env_eval.distributed_mode=none",
+    "rl_games.env_eval.latency.mode=single",
+    "rl_games.env_eval.frameskip=1",
+    "rl_games.env_eval.image_size=224",
+    "rl_games.env_eval.seed=42",
+    "rl_games.env_eval.fixed_episode_seeds=true",
+    "rl_games.env_eval.latency_seed_stride=0",
+    "rl_games.env_eval.task_seed_stride=0",
+    "rl_games.env_eval.latency.values=[0]",
+    "rl_games.env_eval.latency.prompt_map_path=/workspace/starVLA/data/flappy_fix_latency_1/flappy_train__bridge/latency_prompt_map.json",
+    "rl_games.env_eval.mid_train.enabled=true",
+    "rl_games.env_eval.mid_train.interval_steps=100",
+    "rl_games.env_eval.mid_train.num_episodes=5",
+    "rl_games.env_eval.mid_train.max_steps_per_episode=3600",
+    "rl_games.env_eval.mid_train.latencies=[1]",
+    "rl_games.env_eval.post_train.enabled=true",
+    "rl_games.env_eval.post_train.num_episodes=5",
+    "rl_games.env_eval.post_train.max_steps_per_episode=3600",
+    "rl_games.env_eval.post_train.latencies=[0,1,2,3,4]"
+  ],
+  "program": "/workspace/starVLA/starVLA/training/train_starvla_hydra.py",
+  "codePath": "starVLA/training/train_starvla_hydra.py",
+  "codePathLocal": "starVLA/training/train_starvla_hydra.py",
+  "git": {
+    "remote": "https://github.com/talha1503/starVLA.git",
+    "commit": "0bc71a07a07274d35bc9aa60f8d8d14306d45c24"
+  },
+  "email": "dongqianyu99@gmail.com",
+  "root": "/workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/wandb",
+  "host": "2de6a6c642d9",
+  "executable": "/venv/starvla_openvla/bin/python",
+  "cpu_count": 104,
+  "cpu_count_logical": 208,
+  "gpu": "NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition",
+  "gpu_count": 1,
+  "disk": {
+    "/": {
+      "total": "161061273600",
+      "used": "123949047808"
+    }
+  },
+  "memory": {
+    "total": "1081782173696"
+  },
+  "gpu_nvidia": [
+    {
+      "name": "NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition",
+      "memoryTotal": "102641958912",
+      "cudaCores": 24064,
+      "architecture": "Blackwell",
+      "uuid": "GPU-5d139805-8a2f-eab7-b26a-9323f90aa0ff"
+    }
+  ],
+  "cudaVersion": "13.1",
+  "writerId": "44q16xdxvavzayksibbej91m3csw88kd"
+}

wandb/wandb/run-20260607_185700-zhon9num/logs/debug-core.log

@@ -0,0 +1,7 @@
+{"time":"2026-06-07T18:57:00.143815142Z","level":"INFO","msg":"main: starting server","port-filename":"/tmp/tmp20f1s1zq/port-1945.txt","pid":1945,"detached":false,"idle-timeout":600000000000,"log-level":0,"disable-analytics":false,"shutdown-on-parent-exit":false,"enable-dcgm-profiling":false}
+{"time":"2026-06-07T18:57:00.14471536Z","level":"INFO","msg":"server: will exit if parent process dies","ppid":1945}
+{"time":"2026-06-07T18:57:00.14479364Z","level":"INFO","msg":"server: accepting connections","addr":{"Name":"/tmp/wandb-1945-2208-475584196/socket","Net":"unix"}}
+{"time":"2026-06-07T18:57:00.319526172Z","level":"INFO","msg":"connection: ManageConnectionData: new connection created","id":"1(@)"}
+{"time":"2026-06-07T18:57:00.331737225Z","level":"INFO","msg":"handleInformInit: received","streamId":"zhon9num","id":"1(@)"}
+{"time":"2026-06-07T18:57:00.635351579Z","level":"INFO","msg":"handleInformInit: stream started","streamId":"zhon9num","id":"1(@)"}
+{"time":"2026-06-07T18:57:06.226807494Z","level":"INFO","msg":"connection: cancelling request","id":"1(@)","requestId":"0jx2p49wu7uk"}

wandb/wandb/run-20260607_185700-zhon9num/logs/debug-internal.log

@@ -0,0 +1,407 @@
+{"time":"2026-06-07T18:57:00.331870867Z","level":"INFO","msg":"wandb-core"}
+{"time":"2026-06-07T18:57:00.332204327Z","level":"INFO","msg":"stream: starting","core version":"0.27.0"}
+{"time":"2026-06-07T18:57:00.635017625Z","level":"INFO","msg":"stream: created new stream","id":"zhon9num"}
+{"time":"2026-06-07T18:57:00.635153194Z","level":"INFO","msg":"handler: started"}
+{"time":"2026-06-07T18:57:00.635336899Z","level":"INFO","msg":"stream: started"}
+{"time":"2026-06-07T18:57:00.635579128Z","level":"INFO","msg":"writer: started","stream_id":"zhon9num"}
+{"time":"2026-06-07T18:57:00.635683264Z","level":"INFO","msg":"sender: started"}
+{"time":"2026-06-07T18:57:01.219886951Z","level":"INFO","msg":"filestream: sending request","total_files":1,"console_offset":0,"console_lines":2}
+{"time":"2026-06-07T18:57:01.529185153Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T18:57:16.220434943Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":0,"events_lines":1,"console_offset":1,"console_lines":5,"uploaded_len":2}
+{"time":"2026-06-07T18:57:16.47207762Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T18:57:31.219841562Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":1,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T18:57:31.533868633Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T18:57:46.220283515Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":3,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T18:57:46.453574839Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T18:58:01.220129266Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":5,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T18:58:01.469857052Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T18:58:16.220386201Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":7,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T18:58:16.465706204Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T18:58:31.220299175Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":9,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T18:58:31.573517849Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T18:58:46.220402903Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":11,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T18:58:46.436034071Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T18:59:01.220474871Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":13,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T18:59:01.861060339Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T18:59:16.227470747Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":15,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T18:59:16.50098781Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T18:59:31.220618347Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":17,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T18:59:31.436978848Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T18:59:46.220384948Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":19,"events_lines":2,"console_offset":5,"console_lines":2}
+{"time":"2026-06-07T18:59:46.461349785Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:00:01.220037015Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":21,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:00:01.554262631Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:00:16.219861983Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":23,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:00:16.586800845Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:00:31.220140197Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":25,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:00:31.481859546Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:00:46.219873688Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":27,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:00:46.605410951Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:01:01.219841333Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":29,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:01:01.512758545Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:01:16.220568728Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":31,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:01:16.448283191Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:01:31.220015318Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":33,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:01:31.480503069Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:01:46.219816436Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":35,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:01:46.512997861Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:02:01.219961175Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":37,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:02:01.659222647Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:02:16.220021494Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":39,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:02:16.506295079Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:02:31.220299563Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":0,"history_lines":1,"events_offset":41,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:02:31.446191101Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:02:46.220399204Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":43,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:02:46.570917612Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:03:01.219979389Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":45,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:03:01.534120303Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:03:16.219860752Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":47,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:03:16.451648951Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:03:31.219970283Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":49,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:03:31.543155777Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:03:46.219952706Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":51,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:03:46.512862307Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:04:01.219966466Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":53,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:04:01.42031607Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:04:16.220293343Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":55,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:04:16.439842201Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:04:31.220436479Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":57,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:04:31.415504481Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:04:46.219909163Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":59,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:04:46.44764327Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:05:01.219913938Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":61,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:05:01.421995433Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:05:16.220159591Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":1,"history_lines":1,"events_offset":63,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:05:16.432303618Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:05:31.220429705Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":65,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:05:31.440280649Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:05:46.220364842Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":67,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:05:46.420702538Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:06:01.220176773Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":69,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:06:01.431260401Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:06:16.219882241Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":71,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:06:16.427714798Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:06:31.219963614Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":73,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:06:31.881836207Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:06:46.220042677Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":75,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:06:46.415499938Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:07:01.220544921Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":77,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:07:01.460978242Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:07:16.220661418Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":79,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:07:16.430123975Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:07:31.220240828Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":81,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:07:31.425019634Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:07:46.220566907Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":2,"history_lines":1,"events_offset":83,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:07:46.44729853Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:08:01.2200535Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":85,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:08:01.495123709Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:08:16.220073762Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":87,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:08:16.433793808Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:08:31.219847134Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":89,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:08:31.426461883Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:08:46.219828819Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":91,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:08:46.448343657Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:09:01.220241469Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":93,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:09:01.417346871Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:09:16.219969348Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":95,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:09:16.430062383Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:09:31.22024743Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":97,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:09:31.435769991Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:09:46.220152001Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":99,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:09:46.414158501Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:10:01.220499811Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":101,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:10:01.442389279Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:10:16.220434549Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":103,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:10:16.466697454Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:10:31.220110302Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":3,"history_lines":1,"events_offset":105,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:10:31.440175646Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:10:46.22013054Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":107,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:10:46.426986563Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:11:01.219917409Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":109,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:11:01.555997815Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:11:16.219867506Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":111,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:11:16.422670193Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:11:31.220473636Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":113,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:11:31.457477667Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:11:46.220584716Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":115,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:11:46.431197551Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:12:01.219818071Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":117,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:12:01.477558052Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:12:16.220528421Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":119,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:12:16.43536695Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:12:31.220444945Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":121,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:12:31.581195088Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:12:46.220229612Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":123,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:12:46.671750713Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:13:01.230256499Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":125,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:13:01.555727189Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:13:16.220834104Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":4,"history_lines":1,"events_offset":127,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:13:16.444988569Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:13:31.220560704Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":129,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:13:31.803308303Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:13:46.220478118Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":131,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:13:46.434890705Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:14:01.220126216Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":133,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:14:01.452947409Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:14:16.220420501Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":135,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:14:16.43836462Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:14:31.219950685Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":137,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:14:31.419521478Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:14:46.220341724Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":139,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:14:46.415216833Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:15:01.219822892Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":141,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:15:01.436544525Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:15:16.220313642Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":143,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:15:16.422662509Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:15:31.219898008Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":145,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:15:31.426773894Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:15:46.220514719Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":5,"history_lines":1,"events_offset":147,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:15:46.418534834Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:16:01.219786793Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":149,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:16:01.41508833Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:16:16.220017996Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":151,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:16:16.425258633Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:16:31.219879103Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":153,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:16:31.430716753Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:16:46.220148919Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":155,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:16:46.442705143Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:17:01.219987764Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":157,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:17:01.436664143Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:17:16.219995774Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":159,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:17:16.426628306Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:17:31.219760729Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":161,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:17:31.431449457Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:17:46.219837251Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":163,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:17:46.460358139Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:18:01.22010844Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":165,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:18:01.467543266Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:18:16.220168864Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":167,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:18:16.446618439Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:18:31.22029969Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":6,"history_lines":1,"events_offset":169,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:18:31.448948288Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:18:46.220127551Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":171,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:18:46.430535676Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:19:01.220110308Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":173,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:19:01.476461924Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:19:16.21982211Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":175,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:19:16.528908715Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:19:31.219773191Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":177,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:19:31.422441817Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:19:46.220221648Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":179,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:19:46.437204043Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:20:01.220517736Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":181,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:20:01.541137103Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:20:16.220050235Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":183,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:20:16.487837645Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:20:31.220361868Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":185,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:20:31.438540291Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:20:46.22018162Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":187,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:20:46.429937045Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:21:01.220346673Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":7,"history_lines":1,"events_offset":189,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:21:01.438079282Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:21:16.220330063Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":191,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:21:16.428288825Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:21:31.22015474Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":193,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:21:31.42189634Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:21:46.220259181Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":195,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:21:46.423116278Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:22:01.220033912Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":197,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:22:01.429213198Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:22:16.21994841Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":199,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:22:16.497645026Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:22:31.220499376Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":201,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:22:31.569176911Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:22:46.220688043Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":203,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:22:46.515050076Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:23:01.220472363Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":205,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:23:01.534879852Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:23:16.220170985Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":207,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:23:16.479450877Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:23:31.220490465Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":209,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:23:31.469270465Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:23:46.219829253Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":211,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:23:46.439968528Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:24:01.220188951Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":213,"events_lines":2,"console_offset":7,"console_lines":8}
+{"time":"2026-06-07T19:24:01.504413934Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:24:16.220530718Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":215,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:24:16.529658085Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:24:31.220045714Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":217,"events_lines":2,"console_offset":15,"console_lines":8}
+{"time":"2026-06-07T19:24:31.520789019Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:24:46.220084956Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":219,"events_lines":2,"console_offset":22,"console_lines":1}
+{"time":"2026-06-07T19:24:46.513365696Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:25:01.219934464Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":8,"history_lines":1,"events_offset":221,"events_lines":2,"console_offset":5,"console_lines":1}
+{"time":"2026-06-07T19:25:01.562774846Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:25:16.220443276Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":223,"events_lines":2,"console_offset":22,"console_lines":15}
+{"time":"2026-06-07T19:25:16.627579323Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:25:31.220048046Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":225,"events_lines":2,"console_offset":22,"console_lines":1}
+{"time":"2026-06-07T19:25:31.528059206Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:25:46.219996444Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":227,"events_lines":2,"console_offset":22,"console_lines":1}
+{"time":"2026-06-07T19:25:46.480768668Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:26:01.21980408Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":229,"events_lines":2,"console_offset":22,"console_lines":1}
+{"time":"2026-06-07T19:26:01.548576506Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:26:16.220220653Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":231,"events_lines":2,"console_offset":22,"console_lines":1}
+{"time":"2026-06-07T19:26:16.45683673Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:26:31.220107815Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":233,"events_lines":2,"console_offset":22,"console_lines":1}
+{"time":"2026-06-07T19:26:31.568369826Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:26:46.21986403Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":235,"events_lines":2,"console_offset":22,"console_lines":1}
+{"time":"2026-06-07T19:26:46.48470425Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:27:01.219985155Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":237,"events_lines":2,"console_offset":22,"console_lines":1}
+{"time":"2026-06-07T19:27:01.482065076Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:27:16.22067002Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":239,"events_lines":2,"console_offset":22,"console_lines":1}
+{"time":"2026-06-07T19:27:16.449009583Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:27:31.220147701Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":241,"events_lines":2,"console_offset":22,"console_lines":1}
+{"time":"2026-06-07T19:27:31.663291992Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:27:46.220232714Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":243,"events_lines":2,"console_offset":22,"console_lines":1}
+{"time":"2026-06-07T19:27:46.459270815Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:28:01.220054486Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":245,"events_lines":2,"console_offset":22,"console_lines":1}
+{"time":"2026-06-07T19:28:01.427813302Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:28:16.220380442Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":247,"events_lines":2,"console_offset":22,"console_lines":1}
+{"time":"2026-06-07T19:28:16.45497448Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:28:31.220983657Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":9,"history_lines":1,"events_offset":249,"events_lines":2,"console_offset":22,"console_lines":1}
+{"time":"2026-06-07T19:28:31.5259083Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:28:46.220294893Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":251,"events_lines":2,"console_offset":22,"console_lines":1}
+{"time":"2026-06-07T19:28:46.877224775Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:29:01.220136636Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":253,"events_lines":2,"console_offset":22,"console_lines":1}
+{"time":"2026-06-07T19:29:01.515539896Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:29:16.220260265Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":255,"events_lines":2,"console_offset":22,"console_lines":1}
+{"time":"2026-06-07T19:29:16.462758367Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:29:31.220248019Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":257,"events_lines":2,"console_offset":22,"console_lines":1}
+{"time":"2026-06-07T19:29:31.871016544Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:29:46.219897622Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":259,"events_lines":2,"console_offset":22,"console_lines":1}
+{"time":"2026-06-07T19:29:46.462145191Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:30:01.22032218Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":261,"events_lines":2,"console_offset":22,"console_lines":1}
+{"time":"2026-06-07T19:30:01.835244099Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:30:16.220079825Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":263,"events_lines":2,"console_offset":22,"console_lines":1}
+{"time":"2026-06-07T19:30:16.459165916Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:30:31.220517359Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":265,"events_lines":2,"console_offset":22,"console_lines":1}
+{"time":"2026-06-07T19:30:31.464238598Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:30:46.219851365Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":267,"events_lines":2,"console_offset":22,"console_lines":1}
+{"time":"2026-06-07T19:30:46.563150331Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:31:01.220135093Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":269,"events_lines":2,"console_offset":22,"console_lines":1}
+{"time":"2026-06-07T19:31:01.518969069Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:31:16.220806492Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":271,"events_lines":2,"console_offset":22,"console_lines":1}
+{"time":"2026-06-07T19:31:16.579205527Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:31:31.220383443Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":10,"history_lines":1,"events_offset":273,"events_lines":2,"console_offset":22,"console_lines":1}
+{"time":"2026-06-07T19:31:31.4768061Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:31:46.220141857Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":275,"events_lines":2,"console_offset":22,"console_lines":1}
+{"time":"2026-06-07T19:31:46.55151315Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:32:01.219782678Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":277,"events_lines":2,"console_offset":22,"console_lines":1}
+{"time":"2026-06-07T19:32:01.538187805Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:32:16.220530699Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":279,"events_lines":2,"console_offset":22,"console_lines":1}
+{"time":"2026-06-07T19:32:16.76354417Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:32:31.219994669Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":281,"events_lines":2,"console_offset":22,"console_lines":1}
+{"time":"2026-06-07T19:32:31.457438133Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:32:46.220345265Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":283,"events_lines":2,"console_offset":22,"console_lines":1}
+{"time":"2026-06-07T19:32:46.705142126Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:33:01.220442149Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":285,"events_lines":2,"console_offset":22,"console_lines":1}
+{"time":"2026-06-07T19:33:02.114035751Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:33:16.220684619Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":287,"events_lines":2,"console_offset":22,"console_lines":1}
+{"time":"2026-06-07T19:33:16.959562227Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:33:31.220788329Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":289,"events_lines":2,"console_offset":22,"console_lines":1}
+{"time":"2026-06-07T19:33:31.896528151Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:33:46.220083216Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":291,"events_lines":2,"console_offset":22,"console_lines":1}
+{"time":"2026-06-07T19:33:46.45320451Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:34:01.222137148Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":293,"events_lines":2,"console_offset":22,"console_lines":1}
+{"time":"2026-06-07T19:34:02.185073733Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:34:16.22056452Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":295,"events_lines":2,"console_offset":22,"console_lines":1}
+{"time":"2026-06-07T19:34:16.503760954Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:34:31.219928985Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":297,"events_lines":2,"console_offset":22,"console_lines":1}
+{"time":"2026-06-07T19:34:31.473167841Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:34:46.221002631Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":11,"history_lines":1,"events_offset":299,"events_lines":2,"console_offset":22,"console_lines":1}
+{"time":"2026-06-07T19:34:46.529028554Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:35:01.220117377Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":301,"events_lines":2,"console_offset":22,"console_lines":1}
+{"time":"2026-06-07T19:35:01.543488087Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:35:16.22052202Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":303,"events_lines":2,"console_offset":22,"console_lines":1}
+{"time":"2026-06-07T19:35:16.600067558Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:35:31.220811762Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":305,"events_lines":2,"console_offset":22,"console_lines":1}
+{"time":"2026-06-07T19:35:31.544897533Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:35:46.220177768Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":307,"events_lines":2,"console_offset":22,"console_lines":1}
+{"time":"2026-06-07T19:35:46.521285379Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:36:01.220153787Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":309,"events_lines":2,"console_offset":22,"console_lines":1}
+{"time":"2026-06-07T19:36:01.50117921Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:36:16.220155234Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":311,"events_lines":2,"console_offset":22,"console_lines":1}
+{"time":"2026-06-07T19:36:16.45577097Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:36:31.21989688Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":313,"events_lines":2,"console_offset":22,"console_lines":1}
+{"time":"2026-06-07T19:36:31.456797547Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:36:46.21980426Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":315,"events_lines":2,"console_offset":22,"console_lines":1}
+{"time":"2026-06-07T19:36:46.501866945Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:37:01.220537682Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":317,"events_lines":2,"console_offset":22,"console_lines":1}
+{"time":"2026-06-07T19:37:01.511881979Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:37:16.220219842Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":319,"events_lines":2,"console_offset":22,"console_lines":1}
+{"time":"2026-06-07T19:37:16.4547129Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:37:31.219949568Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":321,"events_lines":2,"console_offset":22,"console_lines":1}
+{"time":"2026-06-07T19:37:31.464562085Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:37:46.220782524Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":323,"events_lines":2,"console_offset":22,"console_lines":1}
+{"time":"2026-06-07T19:37:46.492029395Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:38:01.22099069Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":12,"history_lines":1,"events_offset":325,"events_lines":2,"console_offset":22,"console_lines":1}
+{"time":"2026-06-07T19:38:01.468451921Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:38:16.220549173Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":327,"events_lines":2,"console_offset":22,"console_lines":1}
+{"time":"2026-06-07T19:38:16.528692191Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:38:31.22072465Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":329,"events_lines":2,"console_offset":22,"console_lines":1}
+{"time":"2026-06-07T19:38:31.507081973Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:38:46.220861604Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":331,"events_lines":2,"console_offset":22,"console_lines":1}
+{"time":"2026-06-07T19:38:46.503450312Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:39:01.220026763Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":333,"events_lines":2,"console_offset":22,"console_lines":1}
+{"time":"2026-06-07T19:39:01.499230319Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:39:16.219864477Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":335,"events_lines":2,"console_offset":22,"console_lines":1}
+{"time":"2026-06-07T19:39:16.449717438Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:39:31.220122123Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":337,"events_lines":2,"console_offset":22,"console_lines":1}
+{"time":"2026-06-07T19:39:31.46389556Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:39:46.220277533Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":339,"events_lines":2,"console_offset":22,"console_lines":1}
+{"time":"2026-06-07T19:39:46.460822622Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:40:01.220806513Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":341,"events_lines":2,"console_offset":22,"console_lines":1}
+{"time":"2026-06-07T19:40:01.658038849Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:40:16.220473088Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":343,"events_lines":2,"console_offset":22,"console_lines":1}
+{"time":"2026-06-07T19:40:16.62729204Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:40:31.220416823Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":345,"events_lines":2,"console_offset":22,"console_lines":1}
+{"time":"2026-06-07T19:40:31.442552852Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:40:46.220567Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":347,"events_lines":2,"console_offset":22,"console_lines":1}
+{"time":"2026-06-07T19:40:46.670519437Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:41:01.220409301Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":349,"events_lines":2,"console_offset":22,"console_lines":1}
+{"time":"2026-06-07T19:41:01.595661595Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:41:16.22093694Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":13,"history_lines":1,"events_offset":351,"events_lines":2,"console_offset":22,"console_lines":1}
+{"time":"2026-06-07T19:41:16.623164399Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:41:31.220889197Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":353,"events_lines":2,"console_offset":22,"console_lines":1}
+{"time":"2026-06-07T19:41:31.57089624Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:41:46.219862247Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":355,"events_lines":2,"console_offset":22,"console_lines":1}
+{"time":"2026-06-07T19:41:46.506111758Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:42:01.220102158Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":357,"events_lines":2,"console_offset":22,"console_lines":1}
+{"time":"2026-06-07T19:42:01.462755041Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:42:16.220177362Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":359,"events_lines":2,"console_offset":22,"console_lines":1}
+{"time":"2026-06-07T19:42:16.460590304Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:42:31.220147524Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":361,"events_lines":2,"console_offset":22,"console_lines":1}
+{"time":"2026-06-07T19:42:31.454720773Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:42:46.22011352Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":363,"events_lines":2,"console_offset":22,"console_lines":1}
+{"time":"2026-06-07T19:42:46.516698604Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:43:01.220360379Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":365,"events_lines":2,"console_offset":22,"console_lines":1}
+{"time":"2026-06-07T19:43:01.504563105Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:43:16.220023783Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":367,"events_lines":2,"console_offset":22,"console_lines":1}
+{"time":"2026-06-07T19:43:16.522388096Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:43:31.220650641Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":369,"events_lines":2,"console_offset":22,"console_lines":1}
+{"time":"2026-06-07T19:43:31.459235997Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:43:46.220685297Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":371,"events_lines":2,"console_offset":22,"console_lines":1}
+{"time":"2026-06-07T19:43:46.417565528Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:44:01.220450238Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":373,"events_lines":2,"console_offset":22,"console_lines":1}
+{"time":"2026-06-07T19:44:01.439319998Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:44:16.220044723Z","level":"INFO","msg":"filestream: sending request","total_files":4,"history_offset":14,"history_lines":1,"events_offset":375,"events_lines":2,"console_offset":22,"console_lines":1}
+{"time":"2026-06-07T19:44:16.432873937Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:44:31.220178168Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":377,"events_lines":2,"console_offset":22,"console_lines":1}
+{"time":"2026-06-07T19:44:31.447510314Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:44:46.220874582Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":379,"events_lines":2,"console_offset":22,"console_lines":1}
+{"time":"2026-06-07T19:44:46.426868816Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:45:01.220072624Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":381,"events_lines":2,"console_offset":22,"console_lines":1}
+{"time":"2026-06-07T19:45:01.466054662Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:45:16.220140994Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":383,"events_lines":2,"console_offset":22,"console_lines":1}
+{"time":"2026-06-07T19:45:16.408244761Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:45:31.219937195Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":385,"events_lines":2,"console_offset":22,"console_lines":1}
+{"time":"2026-06-07T19:45:31.455896157Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:45:46.219844771Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":387,"events_lines":2,"console_offset":22,"console_lines":1}
+{"time":"2026-06-07T19:45:46.426719636Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:46:01.220477756Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":389,"events_lines":2,"console_offset":22,"console_lines":1}
+{"time":"2026-06-07T19:46:01.487728459Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:46:16.220634346Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":391,"events_lines":2,"console_offset":22,"console_lines":1}
+{"time":"2026-06-07T19:46:16.792334879Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:46:31.220008812Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":393,"events_lines":2,"console_offset":22,"console_lines":1}
+{"time":"2026-06-07T19:46:31.466251267Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-06-07T19:46:46.219799529Z","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":395,"events_lines":2,"console_offset":22,"console_lines":1}
+{"time":"2026-06-07T19:46:46.518353652Z","level":"INFO","msg":"filestream: request sent","status":"200 OK"}

wandb/wandb/run-20260607_185700-zhon9num/logs/debug.log

@@ -0,0 +1,19 @@
+2026-06-07 18:57:00,064 INFO    MainThread:1945 [wandb_setup.py:_flush():81] Current SDK version is 0.27.0
+2026-06-07 18:57:00,064 INFO    MainThread:1945 [wandb_setup.py:_flush():81] Configure stats pid to 1945
+2026-06-07 18:57:00,064 INFO    MainThread:1945 [wandb_setup.py:_flush():81] Loading settings from environment variables
+2026-06-07 18:57:00,064 INFO    MainThread:1945 [wandb_init.py:setup_run_log_directory():723] Logging user logs to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/wandb/wandb/run-20260607_185700-zhon9num/logs/debug.log
+2026-06-07 18:57:00,064 INFO    MainThread:1945 [wandb_init.py:setup_run_log_directory():724] Logging internal logs to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/wandb/wandb/run-20260607_185700-zhon9num/logs/debug-internal.log
+2026-06-07 18:57:00,064 INFO    MainThread:1945 [wandb_init.py:init():851] calling init triggers
+2026-06-07 18:57:00,064 INFO    MainThread:1945 [wandb_init.py:init():856] wandb.init called with sweep_config: {}
+config: {'_wandb': {}}
+2026-06-07 18:57:00,065 INFO    MainThread:1945 [wandb_init.py:init():899] starting backend
+2026-06-07 18:57:00,320 INFO    MainThread:1945 [wandb_init.py:init():914] sending inform_init request
+2026-06-07 18:57:00,636 INFO    MainThread:1945 [wandb_init.py:init():919] backend started and connected
+2026-06-07 18:57:00,639 INFO    MainThread:1945 [wandb_init.py:init():989] updated telemetry
+2026-06-07 18:57:00,647 INFO    MainThread:1945 [wandb_init.py:init():1012] communicating run to backend with 90.0 second timeout
+2026-06-07 18:57:01,137 INFO    MainThread:1945 [wandb_init.py:init():1057] starting run threads in backend
+2026-06-07 18:57:01,215 INFO    MainThread:1945 [wandb_run.py:_console_start():2509] atexit reg
+2026-06-07 18:57:01,215 INFO    MainThread:1945 [wandb_run.py:_redirect():2359] redirect: wrap_raw
+2026-06-07 18:57:01,215 INFO    MainThread:1945 [wandb_run.py:_redirect():2428] Wrapping output streams.
+2026-06-07 18:57:01,215 INFO    MainThread:1945 [wandb_run.py:_redirect():2451] Redirects installed.
+2026-06-07 18:57:01,218 INFO    MainThread:1945 [wandb_init.py:init():1095] run started, returning control to user process

wandb/wandb/run-20260607_185700-zhon9num/run-zhon9num.wandb

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e649b1a3577fbccecb3560903196487fc8871ca115fa32aa7c86d228c284e4d9
+size 1015808

wandb/wandb/run-20260608_022840-lftcj17d/files/config.yaml

@@ -0,0 +1,156 @@
+_wandb:
+    value:
+        cli_version: 0.27.0
+        e:
+            q3pdk48pkts0gntyrorz6d7ezf4isll0:
+                args:
+                    - --config-name
+                    - train
+                    - model=openvla
+                    - env=flappy
+                    - init=bridge
+                    - mode=single
+                    - run_id=openvla_flappy_fix_latency_1
+                    - run_root_dir=/workspace/starVLA/results/Checkpoints
+                    - seed=42
+                    - wandb_entity=dongqianyu99-zhejiang-university
+                    - wandb_project=starVLA_rl_games
+                    - rl_games.env_eval.enabled=true
+                    - checkpoint.sync.enabled=false
+                    - checkpoint.sync.keep_last_n=0
+                    - checkpoint.local.keep_last_n=1
+                    - checkpoint.save_best_model=true
+                    - checkpoint.save_pt_file=false
+                    - trainer.is_resume=true
+                    - trainer.pretrained_checkpoint=/workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_500_state
+                    - trainer.resume_step=500
+                    - trainer.max_train_steps=1500
+                    - trainer.num_warmup_steps=100
+                    - trainer.save_interval=100
+                    - trainer.eval_interval=100
+                    - trainer.eval_num_batches=20
+                    - trainer.logging_frequency=10
+                    - trainer.gradient_accumulation_steps=16
+                    - trainer.distributed_backend=none
+                    - trainer.learning_rate.base=2e-05
+                    - trainer.learning_rate.qwen_vl_interface=1e-05
+                    - trainer.learning_rate.action_model=0.0001
+                    - trainer.lr_scheduler_type=cosine_with_min_lr
+                    - trainer.scheduler_specific_kwargs.min_lr=1e-06
+                    - trainer.loss_scale.vla=1.0
+                    - trainer.loss_scale.vlm=0.1
+                    - trainer.max_grad_norm=1.0
+                    - trainer.weight_decay=0.0
+                    - trainer.gradient_clipping=1.0
+                    - trainer.optimizer.name=AdamW
+                    - trainer.optimizer.betas=[0.9, 0.95]
+                    - trainer.optimizer.eps=1e-08
+                    - trainer.optimizer.weight_decay=1e-08
+                    - trainer.optimizer.fused=false
+                    - trainer.save_format=pt
+                    - framework.name=QwenOFT
+                    - framework.qwenvl.attn_implementation=flash_attention_2
+                    - framework.qwenvl.enable_gradient_checkpointing=true
+                    - framework.action_model.action_dim=7
+                    - framework.action_model.action_env_dim=2
+                    - framework.action_model.state_dim=7
+                    - framework.action_model.loss_type=discrete_ce
+                    - framework.action_model.action_horizon=1
+                    - framework.action_model.future_action_window_size=0
+                    - framework.action_model.past_action_window_size=0
+                    - datasets.vla_data.include_state=true
+                    - datasets.vla_data.action_type=discrete
+                    - datasets.vla_data.sequential_step_sampling=false
+                    - datasets.vla_data.per_device_batch_size=32
+                    - datasets.vla_data.load_all_data_for_training=true
+                    - datasets.vla_data.video_backend=torchvision_av
+                    - datasets.vla_data.data_root_dir=/workspace/starVLA/data/flappy_fix_latency_1
+                    - datasets.vla_data.data_mix=flappy_train__bridge
+                    - datasets.vla_data.eval_data_mix=flappy_train__bridge__val
+                    - framework.qwenvl.base_vlm=/workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct
+                    - rl_games.task=flappy
+                    - rl_games.model_alias=openvla
+                    - rl_games.initialization_mode=bridge
+                    - rl_games.action_carrier=bridge
+                    - rl_games.env_eval.distributed_mode=none
+                    - rl_games.env_eval.latency.mode=single
+                    - rl_games.env_eval.frameskip=1
+                    - rl_games.env_eval.image_size=224
+                    - rl_games.env_eval.seed=42
+                    - rl_games.env_eval.fixed_episode_seeds=true
+                    - rl_games.env_eval.latency_seed_stride=0
+                    - rl_games.env_eval.task_seed_stride=0
+                    - rl_games.env_eval.latency.values=[0]
+                    - rl_games.env_eval.latency.prompt_map_path=/workspace/starVLA/data/flappy_fix_latency_1/flappy_train__bridge/latency_prompt_map.json
+                    - rl_games.env_eval.mid_train.enabled=true
+                    - rl_games.env_eval.mid_train.interval_steps=100
+                    - rl_games.env_eval.mid_train.num_episodes=5
+                    - rl_games.env_eval.mid_train.max_steps_per_episode=3600
+                    - rl_games.env_eval.mid_train.latencies=[1]
+                    - rl_games.env_eval.post_train.enabled=true
+                    - rl_games.env_eval.post_train.num_episodes=5
+                    - rl_games.env_eval.post_train.max_steps_per_episode=3600
+                    - rl_games.env_eval.post_train.latencies=[0,1,2,3,4]
+                codePath: starVLA/training/train_starvla_hydra.py
+                codePathLocal: starVLA/training/train_starvla_hydra.py
+                cpu_count: 104
+                cpu_count_logical: 208
+                cudaVersion: "13.1"
+                disk:
+                    /:
+                        total: "161061273600"
+                        used: "123950485504"
+                email: dongqianyu99@gmail.com
+                executable: /venv/starvla_openvla/bin/python
+                git:
+                    commit: 0bc71a07a07274d35bc9aa60f8d8d14306d45c24
+                    remote: https://github.com/talha1503/starVLA.git
+                gpu: NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition
+                gpu_count: 1
+                gpu_nvidia:
+                    - architecture: Blackwell
+                      cudaCores: 24064
+                      memoryTotal: "102641958912"
+                      name: NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition
+                      uuid: GPU-5d139805-8a2f-eab7-b26a-9323f90aa0ff
+                host: 2de6a6c642d9
+                memory:
+                    total: "1081782173696"
+                os: Linux-6.8.0-90-generic-x86_64-with-glibc2.39
+                program: /workspace/starVLA/starVLA/training/train_starvla_hydra.py
+                python: CPython 3.10.20
+                root: /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/wandb
+                startedAt: "2026-06-08T02:28:40.752684Z"
+                writerId: q3pdk48pkts0gntyrorz6d7ezf4isll0
+        m: []
+        python_version: 3.10.20
+        t:
+            "1":
+                - 1
+                - 11
+                - 41
+                - 49
+                - 50
+                - 63
+                - 71
+                - 80
+                - 83
+            "2":
+                - 1
+                - 11
+                - 41
+                - 49
+                - 50
+                - 63
+                - 71
+                - 80
+                - 83
+            "3":
+                - 2
+                - 13
+                - 61
+            "4": 3.10.20
+            "5": 0.27.0
+            "6": 4.57.0
+            "12": 0.27.0
+            "13": linux-x86_64

wandb/wandb/run-20260608_022840-lftcj17d/files/output.log

@@ -0,0 +1,283 @@
+[2026-06-08 02:28:41,712][starVLA.training.train_starvla][INFO] - ***** Training Configuration *****
+[2026-06-08 02:28:41,712][starVLA.training.train_starvla][INFO] -   Total optimization steps = 1500
+[2026-06-08 02:28:41,712][starVLA.training.train_starvla][INFO] -   Per device batch size = 32
+[2026-06-08 02:28:41,712][starVLA.training.train_starvla][INFO] -   Gradient accumulation steps = 16
+[2026-06-08 02:28:41,712][starVLA.training.train_starvla][INFO] -   Total batch size = 512
+rl-games mid_train eval step=600: 100%|██████████████████████████████████████████████████████| 1/1 [01:20<00:00, 80.51s/it, task=flappy, latency=1]
+[2026-06-08 02:31:22,421][starVLA.training.train_starvla][INFO] - Step 510, Loss: {'train/loss': 0.14785857498645782, 'train/grad_norm_pre_clip': 1.0472160577774048, 'timing/data': 0.00988289900124073, 'timing/model': 1.0121050830930471, 'learning_rate/qwen_vl_interface': 1.0000000000000002e-06, 'learning_rate/action_model': 1e-05, 'epoch': 0.15})
+[2026-06-08 02:34:03,224][starVLA.training.train_starvla][INFO] - Step 520, Loss: {'train/loss': 0.04174887761473656, 'train/grad_norm_pre_clip': 0.5444515943527222, 'timing/data': 0.00963616743683815, 'timing/model': 1.0578863862901926, 'learning_rate/qwen_vl_interface': 2.0000000000000003e-06, 'learning_rate/action_model': 2e-05, 'epoch': 0.15})
+[2026-06-08 02:36:41,917][starVLA.training.train_starvla][INFO] - Step 530, Loss: {'train/loss': 0.11629770696163177, 'train/grad_norm_pre_clip': 1.1295074224472046, 'timing/data': 0.010323092341423035, 'timing/model': 1.055806016549468, 'learning_rate/qwen_vl_interface': 3e-06, 'learning_rate/action_model': 3e-05, 'epoch': 0.16})
+[2026-06-08 02:39:21,995][starVLA.training.train_starvla][INFO] - Step 540, Loss: {'train/loss': 0.15640917420387268, 'train/grad_norm_pre_clip': 1.4473106861114502, 'timing/data': 0.009243069216609001, 'timing/model': 1.0187244340777397, 'learning_rate/qwen_vl_interface': 4.000000000000001e-06, 'learning_rate/action_model': 4e-05, 'epoch': 0.16})
+[2026-06-08 02:41:59,216][starVLA.training.train_starvla][INFO] - Step 550, Loss: {'train/loss': 0.03679528087377548, 'train/grad_norm_pre_clip': 0.8531883955001831, 'timing/data': 0.008215982466936111, 'timing/model': 1.0582027323544025, 'learning_rate/qwen_vl_interface': 5e-06, 'learning_rate/action_model': 5e-05, 'epoch': 0.16})
+[2026-06-08 02:44:38,829][starVLA.training.train_starvla][INFO] - Step 560, Loss: {'train/loss': 0.039581410586833954, 'train/grad_norm_pre_clip': 2.0333027839660645, 'timing/data': 0.01041218638420105, 'timing/model': 1.0539453774690628, 'learning_rate/qwen_vl_interface': 6e-06, 'learning_rate/action_model': 6e-05, 'epoch': 0.17})
+[2026-06-08 02:47:18,238][starVLA.training.train_starvla][INFO] - Step 570, Loss: {'train/loss': 0.048252761363983154, 'train/grad_norm_pre_clip': 0.7267184257507324, 'timing/data': 0.010557536035776138, 'timing/model': 1.0554591342806816, 'learning_rate/qwen_vl_interface': 7e-06, 'learning_rate/action_model': 7e-05, 'epoch': 0.17})
+[2026-06-08 02:49:59,226][starVLA.training.train_starvla][INFO] - Step 580, Loss: {'train/loss': 0.11491948366165161, 'train/grad_norm_pre_clip': 2.2183756828308105, 'timing/data': 0.010399876162409782, 'timing/model': 1.057035032659769, 'learning_rate/qwen_vl_interface': 8.000000000000001e-06, 'learning_rate/action_model': 8e-05, 'epoch': 0.17})
+[2026-06-08 02:52:39,427][starVLA.training.train_starvla][INFO] - Step 590, Loss: {'train/loss': 0.12886273860931396, 'train/grad_norm_pre_clip': 0.30035489797592163, 'timing/data': 0.01050819642841816, 'timing/model': 1.0587196946144104, 'learning_rate/qwen_vl_interface': 9e-06, 'learning_rate/action_model': 9e-05, 'epoch': 0.17})
+[2026-06-08 02:55:45,822] [INFO] [real_accelerator.py:254:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+[2026-06-08 02:55:45,941][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /venv/starvla_openvla/include -fPIC -O2 -isystem /venv/starvla_openvla/include -fPIC -c /tmp/tmp9mzwz8xq/test.c -o /tmp/tmp9mzwz8xq/test.o
+[2026-06-08 02:55:45,967][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat /tmp/tmp9mzwz8xq/test.o -laio -o /tmp/tmp9mzwz8xq/a.out
+[2026-06-08 02:55:46,554][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /venv/starvla_openvla/include -fPIC -O2 -isystem /venv/starvla_openvla/include -fPIC -c /tmp/tmp5by_avfl/test.c -o /tmp/tmp5by_avfl/test.o
+[2026-06-08 02:55:46,581][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat /tmp/tmp5by_avfl/test.o -L/usr/local/cuda -L/usr/local/cuda/lib64 -lcufile -o /tmp/tmp5by_avfl/a.out
+[2026-06-08 02:55:46,638][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /venv/starvla_openvla/include -fPIC -O2 -isystem /venv/starvla_openvla/include -fPIC -c /tmp/tmpxux0dukn/test.c -o /tmp/tmpxux0dukn/test.o
+[2026-06-08 02:55:46,662][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat /tmp/tmpxux0dukn/test.o -laio -o /tmp/tmpxux0dukn/a.out
+ 46%|███████████████████████████████▎                                    | 690/1500 [58:09<4:18:13, 19.13s/it, data_times=0.010, model_times=1.242]
+[2026-06-08 02:57:07,933][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/best_state
+[2026-06-08 02:57:07,939][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading
+[2026-06-08 02:57:13,977][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/best_state/model.safetensors
+[2026-06-08 02:57:32,110][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/best_state/optimizer.bin
+[2026-06-08 02:57:32,110][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/best_state/sampler.bin
+[2026-06-08 02:57:32,111][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/best_state/sampler_1.bin
+[2026-06-08 02:57:32,112][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/best_state/random_states_0.pkl
+✅ New best checkpoint saved at /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/best_state (step=600, score=24.120000)
+[2026-06-08 02:57:32,127][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping best checkpoint upload
+[2026-06-08 02:57:32,127][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 600
+[2026-06-08 02:57:32,128][starVLA.training.train_starvla][INFO] - Step 600, Loss: {'train/loss': 0.04446437954902649, 'train/grad_norm_pre_clip': 2.7018885612487793, 'eval/loss': 0.09853879362344742, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 24.12000000000009, 'rl_games_eval/mid_train/mean_length': 214.4, 'rl_games_eval/mid_train/std_reward': 21.59920368902535, 'rl_games_eval/mid_train/std_length': 174.40137614135963, 'rl_games_eval/mid_train/macro_mean_reward': 24.12000000000009, 'rl_games_eval/mid_train/macro_mean_length': 214.4, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_1/mean_reward': 24.12000000000009, 'rl_games_eval/mid_train/flappy__latency_1/mean_length': 214.4, 'rl_games_eval/mid_train/flappy__latency_1/std_reward': 21.59920368902535, 'rl_games_eval/mid_train/flappy__latency_1/std_length': 174.40137614135963, 'rl_games_eval/mid_train/flappy__latency_1/num_episodes': 5.0, 'checkpoint/current_eval_score': 24.12000000000009, 'checkpoint/best_score': 24.12000000000009, 'checkpoint/best_step': 600.0, 'checkpoint/is_new_best': 1.0, 'timing/data': 0.007813045755028725, 'timing/model': 0.9569010231643915, 'learning_rate/qwen_vl_interface': 1e-05, 'learning_rate/action_model': 0.0001, 'epoch': 0.18})
+[2026-06-08 02:57:32,129][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_600_state
+[2026-06-08 02:57:32,133][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading
+[2026-06-08 02:57:37,971][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_600_state/model.safetensors
+[2026-06-08 02:57:52,790][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_600_state/optimizer.bin
+[2026-06-08 02:57:52,791][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_600_state/sampler.bin
+[2026-06-08 02:57:52,791][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_600_state/sampler_1.bin
+[2026-06-08 02:57:52,792][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_600_state/random_states_0.pkl
+✅ Full training state saved at /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_600_state
+✅ Checkpoint state saved at /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_600_state
+[2026-06-08 02:57:53,156][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 600
+[2026-06-08 02:57:53,156][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration...
+[2026-06-08 02:57:53,171][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved
+[2026-06-08 03:01:04,318][starVLA.training.train_starvla][INFO] - Step 610, Loss: {'train/loss': 0.08884066343307495, 'train/grad_norm_pre_clip': 1.7461992502212524, 'timing/data': 0.00979464128613472, 'timing/model': 1.3480573520064354, 'learning_rate/qwen_vl_interface': 9.998804115976907e-06, 'learning_rate/action_model': 9.998804115976908e-05, 'epoch': 0.18})
+[2026-06-08 03:04:13,622][starVLA.training.train_starvla][INFO] - Step 620, Loss: {'train/loss': 0.2312059849500656, 'train/grad_norm_pre_clip': 2.435190200805664, 'timing/data': 0.010374035686254501, 'timing/model': 1.2583372090011835, 'learning_rate/qwen_vl_interface': 9.995217066071245e-06, 'learning_rate/action_model': 9.995217066071246e-05, 'epoch': 0.18})
+[2026-06-08 03:07:26,103][starVLA.training.train_starvla][INFO] - Step 630, Loss: {'train/loss': 0.21449053287506104, 'train/grad_norm_pre_clip': 2.0557806491851807, 'timing/data': 0.01365184411406517, 'timing/model': 1.2111754715442657, 'learning_rate/qwen_vl_interface': 9.989240656470667e-06, 'learning_rate/action_model': 9.989240656470666e-05, 'epoch': 0.19})
+[2026-06-08 03:10:43,509][starVLA.training.train_starvla][INFO] - Step 640, Loss: {'train/loss': 0.6169573068618774, 'train/grad_norm_pre_clip': 5.35631799697876, 'timing/data': 0.009722961112856865, 'timing/model': 1.438329540193081, 'learning_rate/qwen_vl_interface': 9.980877896477386e-06, 'learning_rate/action_model': 9.980877896477386e-05, 'epoch': 0.19})
+[2026-06-08 03:13:58,031][starVLA.training.train_starvla][INFO] - Step 650, Loss: {'train/loss': 0.04080956056714058, 'train/grad_norm_pre_clip': 0.8581114411354065, 'timing/data': 0.011984208598732948, 'timing/model': 1.261653807014227, 'learning_rate/qwen_vl_interface': 9.970132996992904e-06, 'learning_rate/action_model': 9.970132996992904e-05, 'epoch': 0.19})
+[2026-06-08 03:17:10,730][starVLA.training.train_starvla][INFO] - Step 660, Loss: {'train/loss': 0.18923026323318481, 'train/grad_norm_pre_clip': 0.9931254386901855, 'timing/data': 0.011647291481494904, 'timing/model': 1.260031308978796, 'learning_rate/qwen_vl_interface': 9.957011368397692e-06, 'learning_rate/action_model': 9.95701136839769e-05, 'epoch': 0.2})
+[2026-06-08 03:20:23,256][starVLA.training.train_starvla][INFO] - Step 670, Loss: {'train/loss': 0.09585290402173996, 'train/grad_norm_pre_clip': 1.0940170288085938, 'timing/data': 0.011710893362760544, 'timing/model': 1.2714359946548939, 'learning_rate/qwen_vl_interface': 9.941519617826906e-06, 'learning_rate/action_model': 9.941519617826906e-05, 'epoch': 0.2})
+[2026-06-08 03:23:37,833][starVLA.training.train_starvla][INFO] - Step 680, Loss: {'train/loss': 0.13521376252174377, 'train/grad_norm_pre_clip': 3.671029806137085, 'timing/data': 0.011679081246256828, 'timing/model': 1.559337843209505, 'learning_rate/qwen_vl_interface': 9.923665545843492e-06, 'learning_rate/action_model': 9.923665545843491e-05, 'epoch': 0.2})
+[2026-06-08 03:26:50,794][starVLA.training.train_starvla][INFO] - Step 690, Loss: {'train/loss': 0.062086086720228195, 'train/grad_norm_pre_clip': 1.0403454303741455, 'timing/data': 0.009972494095563889, 'timing/model': 1.2417141627520323, 'learning_rate/qwen_vl_interface': 9.90345814251037e-06, 'learning_rate/action_model': 9.90345814251037e-05, 'epoch': 0.2})
+rl-games mid_train eval step=800: 100%|█| 1/1 [00:31<00:00, 31.86s/it, task=flap0 [1:34:57<3:50:20, 19.74s/it, data_times=0.010, model_times=1.256]
+[2026-06-08 03:31:02,555][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 700
+[2026-06-08 03:31:02,557][starVLA.training.train_starvla][INFO] - Step 700, Loss: {'train/loss': 0.09408611059188843, 'train/grad_norm_pre_clip': 1.4639136791229248, 'eval/loss': 0.1011648178100586, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 9.979999999999983, 'rl_games_eval/mid_train/mean_length': 98.2, 'rl_games_eval/mid_train/std_reward': 3.434763456193156, 'rl_games_eval/mid_train/std_length': 25.43540839066674, 'rl_games_eval/mid_train/macro_mean_reward': 9.979999999999983, 'rl_games_eval/mid_train/macro_mean_length': 98.2, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_1/mean_reward': 9.979999999999983, 'rl_games_eval/mid_train/flappy__latency_1/mean_length': 98.2, 'rl_games_eval/mid_train/flappy__latency_1/std_reward': 3.434763456193156, 'rl_games_eval/mid_train/flappy__latency_1/std_length': 25.43540839066674, 'rl_games_eval/mid_train/flappy__latency_1/num_episodes': 5.0, 'checkpoint/current_eval_score': 9.979999999999983, 'checkpoint/best_score': 24.12000000000009, 'checkpoint/best_step': 600.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.008894853293895721, 'timing/model': 1.1642885133624077, 'learning_rate/qwen_vl_interface': 9.880907582863662e-06, 'learning_rate/action_model': 9.880907582863662e-05, 'epoch': 0.21})
+[2026-06-08 03:31:02,557][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_700_state
+[2026-06-08 03:31:02,561][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading
+[2026-06-08 03:31:08,644][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_700_state/model.safetensors
+[2026-06-08 03:31:22,418][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_700_state/optimizer.bin
+[2026-06-08 03:31:22,419][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_700_state/sampler.bin
+[2026-06-08 03:31:22,419][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_700_state/sampler_1.bin
+[2026-06-08 03:31:22,420][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_700_state/random_states_0.pkl
+✅ Full training state saved at /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_700_state
+✅ Checkpoint state saved at /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_700_state
+[2026-06-08 03:31:22,505][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 700
+[2026-06-08 03:31:22,505][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration...
+[2026-06-08 03:31:22,518][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved
+[2026-06-08 03:34:30,485][starVLA.training.train_starvla][INFO] - Step 710, Loss: {'train/loss': 0.19694210588932037, 'train/grad_norm_pre_clip': 2.3134491443634033, 'timing/data': 0.007648276165127754, 'timing/model': 1.156448507681489, 'learning_rate/qwen_vl_interface': 9.856025221789264e-06, 'learning_rate/action_model': 9.856025221789264e-05, 'epoch': 0.21})
+[2026-06-08 03:37:48,551][starVLA.training.train_starvla][INFO] - Step 720, Loss: {'train/loss': 0.33667615056037903, 'train/grad_norm_pre_clip': 1.7443628311157227, 'timing/data': 0.010220933705568314, 'timing/model': 1.1976180709898472, 'learning_rate/qwen_vl_interface': 9.828823588305305e-06, 'learning_rate/action_model': 9.828823588305303e-05, 'epoch': 0.21})
+[2026-06-08 03:40:57,401][starVLA.training.train_starvla][INFO] - Step 730, Loss: {'train/loss': 0.1062043309211731, 'train/grad_norm_pre_clip': 3.6828112602233887, 'timing/data': 0.009909875690937042, 'timing/model': 1.2409291174262762, 'learning_rate/qwen_vl_interface': 9.79931637925342e-06, 'learning_rate/action_model': 9.79931637925342e-05, 'epoch': 0.22})
+[2026-06-08 03:44:07,118][starVLA.training.train_starvla][INFO] - Step 740, Loss: {'train/loss': 0.1382843255996704, 'train/grad_norm_pre_clip': 1.0776262283325195, 'timing/data': 0.008846268057823181, 'timing/model': 1.2548139616847038, 'learning_rate/qwen_vl_interface': 9.767518452401981e-06, 'learning_rate/action_model': 9.76751845240198e-05, 'epoch': 0.22})
+[2026-06-08 03:47:23,031][starVLA.training.train_starvla][INFO] - Step 750, Loss: {'train/loss': 0.24488884210586548, 'train/grad_norm_pre_clip': 3.646770715713501, 'timing/data': 0.012449927628040314, 'timing/model': 1.2564878407865763, 'learning_rate/qwen_vl_interface': 9.733445818964746e-06, 'learning_rate/action_model': 9.733445818964746e-05, 'epoch': 0.22})
+[2026-06-08 03:50:38,207][starVLA.training.train_starvla][INFO] - Step 760, Loss: {'train/loss': 0.20092175900936127, 'train/grad_norm_pre_clip': 1.4589815139770508, 'timing/data': 0.010395601391792297, 'timing/model': 1.1595692727714777, 'learning_rate/qwen_vl_interface': 9.697115635538753e-06, 'learning_rate/action_model': 9.697115635538753e-05, 'epoch': 0.23})
+[2026-06-08 03:53:54,518][starVLA.training.train_starvla][INFO] - Step 770, Loss: {'train/loss': 0.04230382293462753, 'train/grad_norm_pre_clip': 0.5965043902397156, 'timing/data': 0.01012495718896389, 'timing/model': 1.448968980461359, 'learning_rate/qwen_vl_interface': 9.658546195465436e-06, 'learning_rate/action_model': 9.658546195465435e-05, 'epoch': 0.23})
+[2026-06-08 03:57:06,624][starVLA.training.train_starvla][INFO] - Step 780, Loss: {'train/loss': 0.12880125641822815, 'train/grad_norm_pre_clip': 3.2280399799346924, 'timing/data': 0.011633196845650673, 'timing/model': 1.2479083891957998, 'learning_rate/qwen_vl_interface': 9.617756919619393e-06, 'learning_rate/action_model': 9.617756919619392e-05, 'epoch': 0.23})
+[2026-06-08 04:00:24,321][starVLA.training.train_starvla][INFO] - Step 790, Loss: {'train/loss': 0.22798243165016174, 'train/grad_norm_pre_clip': 2.152913808822632, 'timing/data': 0.010926170274615288, 'timing/model': 1.2523880619555712, 'learning_rate/qwen_vl_interface': 9.574768346629376e-06, 'learning_rate/action_model': 9.574768346629376e-05, 'epoch': 0.23})
+rl-games mid_train eval step=900: 100%|█| 1/1 [00:46<00:00, 46.24s/it, task=flap0 [2:08:29<3:09:06, 18.91s/it, data_times=0.010, model_times=1.256]
+[2026-06-08 04:04:36,030][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 800
+[2026-06-08 04:04:36,031][starVLA.training.train_starvla][INFO] - Step 800, Loss: {'train/loss': 0.026743575930595398, 'train/grad_norm_pre_clip': 1.0122240781784058, 'eval/loss': 0.10519333928823471, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 7.699999999999989, 'rl_games_eval/mid_train/mean_length': 80.8, 'rl_games_eval/mid_train/std_reward': 4.08215629294126, 'rl_games_eval/mid_train/std_length': 30.32754523531372, 'rl_games_eval/mid_train/macro_mean_reward': 7.699999999999989, 'rl_games_eval/mid_train/macro_mean_length': 80.8, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_1/mean_reward': 7.699999999999989, 'rl_games_eval/mid_train/flappy__latency_1/mean_length': 80.8, 'rl_games_eval/mid_train/flappy__latency_1/std_reward': 4.08215629294126, 'rl_games_eval/mid_train/flappy__latency_1/std_length': 30.32754523531372, 'rl_games_eval/mid_train/flappy__latency_1/num_episodes': 5.0, 'checkpoint/current_eval_score': 7.699999999999989, 'checkpoint/best_score': 24.12000000000009, 'checkpoint/best_step': 600.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.010297557339072227, 'timing/model': 1.2559085674583912, 'learning_rate/qwen_vl_interface': 9.52960212253649e-06, 'learning_rate/action_model': 9.529602122536491e-05, 'epoch': 0.24})
+[2026-06-08 04:04:36,032][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_800_state
+[2026-06-08 04:04:36,035][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading
+[2026-06-08 04:04:42,337][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_800_state/model.safetensors
+[2026-06-08 04:04:55,707][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_800_state/optimizer.bin
+[2026-06-08 04:04:55,707][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_800_state/sampler.bin
+[2026-06-08 04:04:55,707][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_800_state/sampler_1.bin
+[2026-06-08 04:04:55,708][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_800_state/random_states_0.pkl
+✅ Full training state saved at /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_800_state
+✅ Checkpoint state saved at /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_800_state
+[2026-06-08 04:04:55,803][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 800
+[2026-06-08 04:04:55,803][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration...
+[2026-06-08 04:04:55,815][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved
+[2026-06-08 04:08:03,431][starVLA.training.train_starvla][INFO] - Step 810, Loss: {'train/loss': 0.08235027641057968, 'train/grad_norm_pre_clip': 0.8081689476966858, 'timing/data': 0.010712772607803345, 'timing/model': 1.2615945506840944, 'learning_rate/qwen_vl_interface': 9.482280989894748e-06, 'learning_rate/action_model': 9.482280989894748e-05, 'epoch': 0.24})
+[2026-06-08 04:11:17,626][starVLA.training.train_starvla][INFO] - Step 820, Loss: {'train/loss': 0.11566474288702011, 'train/grad_norm_pre_clip': 1.46780526638031, 'timing/data': 0.011909419670701027, 'timing/model': 1.2576640285551548, 'learning_rate/qwen_vl_interface': 9.432828776319507e-06, 'learning_rate/action_model': 9.432828776319505e-05, 'epoch': 0.24})
+[2026-06-08 04:14:32,121][starVLA.training.train_starvla][INFO] - Step 830, Loss: {'train/loss': 0.22728708386421204, 'train/grad_norm_pre_clip': 1.0950161218643188, 'timing/data': 0.010704748332500458, 'timing/model': 1.2469014767557383, 'learning_rate/qwen_vl_interface': 9.381270382489554e-06, 'learning_rate/action_model': 9.381270382489554e-05, 'epoch': 0.25})
+[2026-06-08 04:17:43,623][starVLA.training.train_starvla][INFO] - Step 840, Loss: {'train/loss': 0.0630272850394249, 'train/grad_norm_pre_clip': 0.7764682769775391, 'timing/data': 0.010154886171221733, 'timing/model': 1.2599844448268414, 'learning_rate/qwen_vl_interface': 9.327631769608864e-06, 'learning_rate/action_model': 9.327631769608864e-05, 'epoch': 0.25})
+[2026-06-08 04:20:56,829][starVLA.training.train_starvla][INFO] - Step 850, Loss: {'train/loss': 0.08732070028781891, 'train/grad_norm_pre_clip': 0.467477023601532, 'timing/data': 0.010380668565630913, 'timing/model': 1.2687833160161972, 'learning_rate/qwen_vl_interface': 9.27193994633435e-06, 'learning_rate/action_model': 9.27193994633435e-05, 'epoch': 0.25})
+[2026-06-08 04:24:11,043][starVLA.training.train_starvla][INFO] - Step 860, Loss: {'train/loss': 0.199907124042511, 'train/grad_norm_pre_clip': 0.7404685616493225, 'timing/data': 0.00981375016272068, 'timing/model': 1.3720009662210941, 'learning_rate/qwen_vl_interface': 9.214222955176189e-06, 'learning_rate/action_model': 9.214222955176189e-05, 'epoch': 0.25})
+[2026-06-08 04:27:29,829][starVLA.training.train_starvla][INFO] - Step 870, Loss: {'train/loss': 0.12482769042253494, 'train/grad_norm_pre_clip': 0.583684504032135, 'timing/data': 0.012422095984220505, 'timing/model': 1.2570643611252308, 'learning_rate/qwen_vl_interface': 9.154509858377574e-06, 'learning_rate/action_model': 9.154509858377574e-05, 'epoch': 0.26})
+[2026-06-08 04:30:45,417][starVLA.training.train_starvla][INFO] - Step 880, Loss: {'train/loss': 0.16042155027389526, 'train/grad_norm_pre_clip': 0.7780199646949768, 'timing/data': 0.013919441029429436, 'timing/model': 1.3473796378821135, 'learning_rate/qwen_vl_interface': 9.092830723281e-06, 'learning_rate/action_model': 9.092830723281e-05, 'epoch': 0.26})
+[2026-06-08 04:34:00,736][starVLA.training.train_starvla][INFO] - Step 890, Loss: {'train/loss': 0.24358049035072327, 'train/grad_norm_pre_clip': 1.528757929801941, 'timing/data': 0.010966861620545387, 'timing/model': 1.1853130254894495, 'learning_rate/qwen_vl_interface': 9.02921660718846e-06, 'learning_rate/action_model': 9.029216607188459e-05, 'epoch': 0.26})
+rl-games mid_train eval step=1000: 100%|█| 1/1 [00:31<00:00, 31.44s/it, task=fla0 [2:42:04<2:33:29, 18.42s/it, data_times=0.009, model_times=1.165]
+[2026-06-08 04:38:23,716][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 900
+[2026-06-08 04:38:23,718][starVLA.training.train_starvla][INFO] - Step 900, Loss: {'train/loss': 0.10494193434715271, 'train/grad_norm_pre_clip': 1.7167644500732422, 'eval/loss': 0.10630352795124054, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 13.860000000000031, 'rl_games_eval/mid_train/mean_length': 129.8, 'rl_games_eval/mid_train/std_reward': 10.23241906882245, 'rl_games_eval/mid_train/std_length': 83.25959404176795, 'rl_games_eval/mid_train/macro_mean_reward': 13.860000000000031, 'rl_games_eval/mid_train/macro_mean_length': 129.8, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_1/mean_reward': 13.860000000000031, 'rl_games_eval/mid_train/flappy__latency_1/mean_length': 129.8, 'rl_games_eval/mid_train/flappy__latency_1/std_reward': 10.23241906882245, 'rl_games_eval/mid_train/flappy__latency_1/std_length': 83.25959404176795, 'rl_games_eval/mid_train/flappy__latency_1/num_episodes': 5.0, 'checkpoint/current_eval_score': 13.860000000000031, 'checkpoint/best_score': 24.12000000000009, 'checkpoint/best_step': 600.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.010333351790904999, 'timing/model': 1.256254507228732, 'learning_rate/qwen_vl_interface': 8.963699541723143e-06, 'learning_rate/action_model': 8.963699541723142e-05, 'epoch': 0.27})
+[2026-06-08 04:38:23,719][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_900_state
+[2026-06-08 04:38:23,722][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading
+[2026-06-08 04:38:29,894][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_900_state/model.safetensors
+[2026-06-08 04:38:43,706][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_900_state/optimizer.bin
+[2026-06-08 04:38:43,707][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_900_state/sampler.bin
+[2026-06-08 04:38:43,707][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_900_state/sampler_1.bin
+[2026-06-08 04:38:43,708][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_900_state/random_states_0.pkl
+✅ Full training state saved at /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_900_state
+✅ Checkpoint state saved at /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_900_state
+[2026-06-08 04:38:43,798][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 900
+[2026-06-08 04:38:43,798][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration...
+[2026-06-08 04:38:43,811][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved
+[2026-06-08 04:41:50,410][starVLA.training.train_starvla][INFO] - Step 910, Loss: {'train/loss': 0.1504046618938446, 'train/grad_norm_pre_clip': 1.7034003734588623, 'timing/data': 0.01012641191482544, 'timing/model': 1.2499719206243753, 'learning_rate/qwen_vl_interface': 8.896312516700553e-06, 'learning_rate/action_model': 8.896312516700554e-05, 'epoch': 0.27})
+[2026-06-08 04:45:05,433][starVLA.training.train_starvla][INFO] - Step 920, Loss: {'train/loss': 0.026497211307287216, 'train/grad_norm_pre_clip': 0.28330326080322266, 'timing/data': 0.010129151865839958, 'timing/model': 1.3549129571765661, 'learning_rate/qwen_vl_interface': 8.827089463517153e-06, 'learning_rate/action_model': 8.827089463517153e-05, 'epoch': 0.27})
+[2026-06-08 04:48:24,431][starVLA.training.train_starvla][INFO] - Step 930, Loss: {'train/loss': 0.1441197246313095, 'train/grad_norm_pre_clip': 0.888932466506958, 'timing/data': 0.0006872452795505524, 'timing/model': 1.232828725129366, 'learning_rate/qwen_vl_interface': 8.75606523806487e-06, 'learning_rate/action_model': 8.756065238064868e-05, 'epoch': 0.28})
+[2026-06-08 04:51:41,320][starVLA.training.train_starvla][INFO] - Step 940, Loss: {'train/loss': 0.06046869978308678, 'train/grad_norm_pre_clip': 0.6223641633987427, 'timing/data': 0.0006999224424362183, 'timing/model': 1.2607584223151207, 'learning_rate/qwen_vl_interface': 8.68327560318011e-06, 'learning_rate/action_model': 8.68327560318011e-05, 'epoch': 0.28})
+[2026-06-08 04:54:56,115][starVLA.training.train_starvla][INFO] - Step 950, Loss: {'train/loss': 0.03481486812233925, 'train/grad_norm_pre_clip': 0.8553818464279175, 'timing/data': 0.001046663150191307, 'timing/model': 1.2562530543655157, 'learning_rate/qwen_vl_interface': 8.608757210636101e-06, 'learning_rate/action_model': 8.608757210636101e-05, 'epoch': 0.28})
+[2026-06-08 04:58:05,620][starVLA.training.train_starvla][INFO] - Step 960, Loss: {'train/loss': 0.22162754833698273, 'train/grad_norm_pre_clip': 4.34283971786499, 'timing/data': 0.0006981156766414642, 'timing/model': 1.260053401812911, 'learning_rate/qwen_vl_interface': 8.532547582687609e-06, 'learning_rate/action_model': 8.532547582687608e-05, 'epoch': 0.28})
+[2026-06-08 05:01:23,328][starVLA.training.train_starvla][INFO] - Step 970, Loss: {'train/loss': 0.06828710436820984, 'train/grad_norm_pre_clip': 0.6473269462585449, 'timing/data': 0.011569458991289139, 'timing/model': 1.256771007552743, 'learning_rate/qwen_vl_interface': 8.454685093177356e-06, 'learning_rate/action_model': 8.454685093177354e-05, 'epoch': 0.29})
+[2026-06-08 05:04:36,827][starVLA.training.train_starvla][INFO] - Step 980, Loss: {'train/loss': 0.06886345148086548, 'train/grad_norm_pre_clip': 1.0889666080474854, 'timing/data': 0.010370265692472458, 'timing/model': 1.2603094186633825, 'learning_rate/qwen_vl_interface': 8.375208948213636e-06, 'learning_rate/action_model': 8.375208948213635e-05, 'epoch': 0.29})
+[2026-06-08 05:07:39,058][starVLA.training.train_starvla][INFO] - Step 990, Loss: {'train/loss': 0.08339225500822067, 'train/grad_norm_pre_clip': 1.1716196537017822, 'timing/data': 0.010570503771305084, 'timing/model': 1.3141696825623512, 'learning_rate/qwen_vl_interface': 8.294159166428864e-06, 'learning_rate/action_model': 8.294159166428864e-05, 'epoch': 0.29})
+rl-games mid_train eval step=1100: 100%|█████████████████████████████████████████████████████| 1/1 [01:22<00:00, 82.28s/it, task=flappy, latency=1]
+[2026-06-08 05:11:41,629][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1000
+[2026-06-08 05:11:41,632][starVLA.training.train_starvla][INFO] - Step 1000, Loss: {'train/loss': 0.043065138161182404, 'train/grad_norm_pre_clip': 0.9228273034095764, 'eval/loss': 0.10095208138227463, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 8.079999999999988, 'rl_games_eval/mid_train/mean_length': 82.8, 'rl_games_eval/mid_train/std_reward': 1.4218298069740927, 'rl_games_eval/mid_train/std_length': 14.218298069740976, 'rl_games_eval/mid_train/macro_mean_reward': 8.079999999999988, 'rl_games_eval/mid_train/macro_mean_length': 82.8, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_1/mean_reward': 8.079999999999988, 'rl_games_eval/mid_train/flappy__latency_1/mean_length': 82.8, 'rl_games_eval/mid_train/flappy__latency_1/std_reward': 1.4218298069740927, 'rl_games_eval/mid_train/flappy__latency_1/std_length': 14.218298069740976, 'rl_games_eval/mid_train/flappy__latency_1/num_episodes': 5.0, 'checkpoint/current_eval_score': 8.079999999999988, 'checkpoint/best_score': 24.12000000000009, 'checkpoint/best_step': 600.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.008860763162374496, 'timing/model': 1.164587363600731, 'learning_rate/qwen_vl_interface': 8.211576558828985e-06, 'learning_rate/action_model': 8.211576558828985e-05, 'epoch': 0.3})
+[2026-06-08 05:11:41,633][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_1000_state
+[2026-06-08 05:11:41,638][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading
+[2026-06-08 05:11:49,263][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_1000_state/model.safetensors
+[2026-06-08 05:12:06,718][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_1000_state/optimizer.bin
+[2026-06-08 05:12:06,718][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_1000_state/sampler.bin
+[2026-06-08 05:12:06,718][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_1000_state/sampler_1.bin
+[2026-06-08 05:12:06,720][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_1000_state/random_states_0.pkl
+✅ Full training state saved at /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_1000_state
+✅ Checkpoint state saved at /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_1000_state
+[2026-06-08 05:12:06,813][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1000
+[2026-06-08 05:12:06,814][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration...
+[2026-06-08 05:12:06,826][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved
+[2026-06-08 05:15:17,808][starVLA.training.train_starvla][INFO] - Step 1010, Loss: {'train/loss': 0.06643424928188324, 'train/grad_norm_pre_clip': 1.9465690851211548, 'timing/data': 0.010009264573454857, 'timing/model': 1.21424631588161, 'learning_rate/qwen_vl_interface': 8.127502708243895e-06, 'learning_rate/action_model': 8.127502708243895e-05, 'epoch': 0.3})
+[2026-06-08 05:18:32,615][starVLA.training.train_starvla][INFO] - Step 1020, Loss: {'train/loss': 0.021961456164717674, 'train/grad_norm_pre_clip': 0.5190539956092834, 'timing/data': 0.010486900806427002, 'timing/model': 1.248643545433879, 'learning_rate/qwen_vl_interface': 8.041979948389248e-06, 'learning_rate/action_model': 8.041979948389248e-05, 'epoch': 0.3})
+[2026-06-08 05:21:43,527][starVLA.training.train_starvla][INFO] - Step 1030, Loss: {'train/loss': 0.11902540177106857, 'train/grad_norm_pre_clip': 2.2872815132141113, 'timing/data': 0.009541437029838562, 'timing/model': 1.2074501030147076, 'learning_rate/qwen_vl_interface': 7.955051342550148e-06, 'learning_rate/action_model': 7.955051342550148e-05, 'epoch': 0.31})
+[2026-06-08 05:24:52,218][starVLA.training.train_starvla][INFO] - Step 1040, Loss: {'train/loss': 0.11230383813381195, 'train/grad_norm_pre_clip': 0.641843318939209, 'timing/data': 0.010753313079476357, 'timing/model': 1.2581883277744055, 'learning_rate/qwen_vl_interface': 7.866760661897487e-06, 'learning_rate/action_model': 7.866760661897487e-05, 'epoch': 0.31})
+[2026-06-08 05:28:06,630][starVLA.training.train_starvla][INFO] - Step 1050, Loss: {'train/loss': 0.07386970520019531, 'train/grad_norm_pre_clip': 2.0325117111206055, 'timing/data': 0.011218462139368057, 'timing/model': 1.461636632680893, 'learning_rate/qwen_vl_interface': 7.777152363447849e-06, 'learning_rate/action_model': 7.77715236344785e-05, 'epoch': 0.31})
+[2026-06-08 05:31:16,635][starVLA.training.train_starvla][INFO] - Step 1060, Loss: {'train/loss': 0.11535225063562393, 'train/grad_norm_pre_clip': 0.9882320165634155, 'timing/data': 0.01395397074520588, 'timing/model': 1.34351110085845, 'learning_rate/qwen_vl_interface': 7.686271567678056e-06, 'learning_rate/action_model': 7.686271567678054e-05, 'epoch': 0.31})
+[2026-06-08 05:34:24,092][starVLA.training.train_starvla][INFO] - Step 1070, Loss: {'train/loss': 0.0681602731347084, 'train/grad_norm_pre_clip': 1.3164373636245728, 'timing/data': 0.009659260511398315, 'timing/model': 1.1513372603803873, 'learning_rate/qwen_vl_interface': 7.594164035805648e-06, 'learning_rate/action_model': 7.594164035805648e-05, 'epoch': 0.32})
+[2026-06-08 05:37:35,824][starVLA.training.train_starvla][INFO] - Step 1080, Loss: {'train/loss': 0.03250638768076897, 'train/grad_norm_pre_clip': 0.5309199094772339, 'timing/data': 0.010097391903400421, 'timing/model': 1.2179136835038662, 'learning_rate/qwen_vl_interface': 7.5008761467467435e-06, 'learning_rate/action_model': 7.500876146746743e-05, 'epoch': 0.32})
+[2026-06-08 05:40:41,721][starVLA.training.train_starvla][INFO] - Step 1090, Loss: {'train/loss': 0.12813404202461243, 'train/grad_norm_pre_clip': 3.2457115650177, 'timing/data': 0.009411433711647987, 'timing/model': 1.2594144362956285, 'learning_rate/qwen_vl_interface': 7.406454873762848e-06, 'learning_rate/action_model': 7.406454873762848e-05, 'epoch': 0.32})
+ 79%|███████████████████████████████████████████████████▌             | 1190/1500 [3:46:37<1:40:15, 19.41s/it, data_times=0.010, model_times=1.457]
+[2026-06-08 05:45:39,647][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/best_state
+[2026-06-08 05:45:39,651][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading
+[2026-06-08 05:45:45,607][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/best_state/model.safetensors
+[2026-06-08 05:46:00,921][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/best_state/optimizer.bin
+[2026-06-08 05:46:00,921][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/best_state/sampler.bin
+[2026-06-08 05:46:00,922][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/best_state/sampler_1.bin
+[2026-06-08 05:46:00,923][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/best_state/random_states_0.pkl
+✅ New best checkpoint saved at /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/best_state (step=1100, score=24.940000)
+[2026-06-08 05:46:00,939][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping best checkpoint upload
+[2026-06-08 05:46:00,939][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1100
+[2026-06-08 05:46:00,941][starVLA.training.train_starvla][INFO] - Step 1100, Loss: {'train/loss': 0.04780591279268265, 'train/grad_norm_pre_clip': 0.7340912818908691, 'eval/loss': 0.07767148315906525, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 24.940000000000126, 'rl_games_eval/mid_train/mean_length': 219.0, 'rl_games_eval/mid_train/std_reward': 16.381282001113483, 'rl_games_eval/mid_train/std_length': 133.51404420509476, 'rl_games_eval/mid_train/macro_mean_reward': 24.940000000000126, 'rl_games_eval/mid_train/macro_mean_length': 219.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_1/mean_reward': 24.940000000000126, 'rl_games_eval/mid_train/flappy__latency_1/mean_length': 219.0, 'rl_games_eval/mid_train/flappy__latency_1/std_reward': 16.381282001113483, 'rl_games_eval/mid_train/flappy__latency_1/std_length': 133.51404420509476, 'rl_games_eval/mid_train/flappy__latency_1/num_episodes': 5.0, 'checkpoint/current_eval_score': 24.940000000000126, 'checkpoint/best_score': 24.940000000000126, 'checkpoint/best_step': 1100.0, 'checkpoint/is_new_best': 1.0, 'timing/data': 0.011445408686995506, 'timing/model': 1.2202512193471193, 'learning_rate/qwen_vl_interface': 7.310947760808401e-06, 'learning_rate/action_model': 7.3109477608084e-05, 'epoch': 0.33})
+[2026-06-08 05:46:00,942][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_1100_state
+[2026-06-08 05:46:00,953][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading
+[2026-06-08 05:46:08,864][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_1100_state/model.safetensors
+[2026-06-08 05:46:23,152][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_1100_state/optimizer.bin
+[2026-06-08 05:46:23,153][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_1100_state/sampler.bin
+[2026-06-08 05:46:23,153][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_1100_state/sampler_1.bin
+[2026-06-08 05:46:23,154][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_1100_state/random_states_0.pkl
+✅ Full training state saved at /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_1100_state
+✅ Checkpoint state saved at /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_1100_state
+[2026-06-08 05:46:23,251][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1100
+[2026-06-08 05:46:23,251][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration...
+[2026-06-08 05:46:23,265][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved
+[2026-06-08 05:49:31,523][starVLA.training.train_starvla][INFO] - Step 1110, Loss: {'train/loss': 0.16532175242900848, 'train/grad_norm_pre_clip': 2.9323229789733887, 'timing/data': 0.011479422450065613, 'timing/model': 1.2202869094908237, 'learning_rate/qwen_vl_interface': 7.214402898590974e-06, 'learning_rate/action_model': 7.214402898590973e-05, 'epoch': 0.33})
+[2026-06-08 05:52:41,114][starVLA.training.train_starvla][INFO] - Step 1120, Loss: {'train/loss': 0.16425183415412903, 'train/grad_norm_pre_clip': 2.0046119689941406, 'timing/data': 0.012083834037184715, 'timing/model': 1.254353767260909, 'learning_rate/qwen_vl_interface': 7.116868900356138e-06, 'learning_rate/action_model': 7.116868900356138e-05, 'epoch': 0.33})
+[2026-06-08 05:55:53,619][starVLA.training.train_starvla][INFO] - Step 1130, Loss: {'train/loss': 0.044812172651290894, 'train/grad_norm_pre_clip': 0.5543600916862488, 'timing/data': 0.011260369792580605, 'timing/model': 1.2566115446388721, 'learning_rate/qwen_vl_interface': 7.01839487740924e-06, 'learning_rate/action_model': 7.01839487740924e-05, 'epoch': 0.33})
+[2026-06-08 05:59:14,286][starVLA.training.train_starvla][INFO] - Step 1140, Loss: {'train/loss': 0.01742750033736229, 'train/grad_norm_pre_clip': 0.8612880110740662, 'timing/data': 0.013038020581007004, 'timing/model': 1.5489813480526209, 'learning_rate/qwen_vl_interface': 6.919030414386378e-06, 'learning_rate/action_model': 6.919030414386379e-05, 'epoch': 0.34})
+[2026-06-08 06:02:30,403][starVLA.training.train_starvla][INFO] - Step 1150, Loss: {'train/loss': 0.04407715052366257, 'train/grad_norm_pre_clip': 1.12485671043396, 'timing/data': 0.009085983037948608, 'timing/model': 1.1555767077952623, 'learning_rate/qwen_vl_interface': 6.818825544287045e-06, 'learning_rate/action_model': 6.818825544287045e-05, 'epoch': 0.34})
+[2026-06-08 06:05:41,329][starVLA.training.train_starvla][INFO] - Step 1160, Loss: {'train/loss': 0.0421944260597229, 'train/grad_norm_pre_clip': 0.7835307717323303, 'timing/data': 0.010449636727571487, 'timing/model': 1.216680582612753, 'learning_rate/qwen_vl_interface': 6.7178307232810015e-06, 'learning_rate/action_model': 6.717830723281e-05, 'epoch': 0.34})
+[2026-06-08 06:08:58,941][starVLA.training.train_starvla][INFO] - Step 1170, Loss: {'train/loss': 0.04635126143693924, 'train/grad_norm_pre_clip': 0.3387547731399536, 'timing/data': 0.011028308421373367, 'timing/model': 1.2670236714184284, 'learning_rate/qwen_vl_interface': 6.616096805302083e-06, 'learning_rate/action_model': 6.616096805302082e-05, 'epoch': 0.35})
+[2026-06-08 06:12:08,595][starVLA.training.train_starvla][INFO] - Step 1180, Loss: {'train/loss': 0.10106848180294037, 'train/grad_norm_pre_clip': 1.9791182279586792, 'timing/data': 0.008970709517598152, 'timing/model': 1.1509433034807444, 'learning_rate/qwen_vl_interface': 6.513675016441707e-06, 'learning_rate/action_model': 6.513675016441707e-05, 'epoch': 0.35})
+[2026-06-08 06:15:18,727][starVLA.training.train_starvla][INFO] - Step 1190, Loss: {'train/loss': 0.031249359250068665, 'train/grad_norm_pre_clip': 0.2567000091075897, 'timing/data': 0.010415764525532722, 'timing/model': 1.45666971988976, 'learning_rate/qwen_vl_interface': 6.410616929154997e-06, 'learning_rate/action_model': 6.410616929154996e-05, 'epoch': 0.35})
+rl-games mid_train eval step=1300: 100%|█████████████████████████████████████████████████████| 1/1 [00:43<00:00, 43.75s/it, task=flappy, latency=1]
+[2026-06-08 06:19:44,671][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1200
+[2026-06-08 06:19:44,672][starVLA.training.train_starvla][INFO] - Step 1200, Loss: {'train/loss': 0.059253789484500885, 'train/grad_norm_pre_clip': 0.5499390959739685, 'eval/loss': 0.09212758392095566, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 17.100000000000037, 'rl_games_eval/mid_train/mean_length': 156.8, 'rl_games_eval/mid_train/std_reward': 26.150028680672683, 'rl_games_eval/mid_train/std_length': 211.10035528155797, 'rl_games_eval/mid_train/macro_mean_reward': 17.100000000000037, 'rl_games_eval/mid_train/macro_mean_length': 156.8, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_1/mean_reward': 17.100000000000037, 'rl_games_eval/mid_train/flappy__latency_1/mean_length': 156.8, 'rl_games_eval/mid_train/flappy__latency_1/std_reward': 26.150028680672683, 'rl_games_eval/mid_train/flappy__latency_1/std_length': 211.10035528155797, 'rl_games_eval/mid_train/flappy__latency_1/num_episodes': 5.0, 'checkpoint/current_eval_score': 17.100000000000037, 'checkpoint/best_score': 24.940000000000126, 'checkpoint/best_step': 1100.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.009885326027870178, 'timing/model': 1.1901550088077784, 'learning_rate/qwen_vl_interface': 6.306974436292493e-06, 'learning_rate/action_model': 6.306974436292494e-05, 'epoch': 0.36})
+[2026-06-08 06:19:44,673][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_1200_state
+[2026-06-08 06:19:44,677][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading
+[2026-06-08 06:19:50,690][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_1200_state/model.safetensors
+[2026-06-08 06:20:04,743][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_1200_state/optimizer.bin
+[2026-06-08 06:20:04,744][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_1200_state/sampler.bin
+[2026-06-08 06:20:04,744][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_1200_state/sampler_1.bin
+[2026-06-08 06:20:04,745][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_1200_state/random_states_0.pkl
+✅ Full training state saved at /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_1200_state
+✅ Checkpoint state saved at /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_1200_state
+[2026-06-08 06:20:04,846][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1200
+[2026-06-08 06:20:04,846][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration...
+[2026-06-08 06:20:04,859][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved
+[2026-06-08 06:23:11,092][starVLA.training.train_starvla][INFO] - Step 1210, Loss: {'train/loss': 0.168832466006279, 'train/grad_norm_pre_clip': 1.3696372509002686, 'timing/data': 0.009317239746451378, 'timing/model': 1.1600620280951262, 'learning_rate/qwen_vl_interface': 6.202799724970546e-06, 'learning_rate/action_model': 6.202799724970546e-05, 'epoch': 0.36})
+[2026-06-08 06:26:20,640][starVLA.training.train_starvla][INFO] - Step 1220, Loss: {'train/loss': 0.11728746443986893, 'train/grad_norm_pre_clip': 0.764240562915802, 'timing/data': 0.010577192530035973, 'timing/model': 1.3747284729033709, 'learning_rate/qwen_vl_interface': 6.098145250293525e-06, 'learning_rate/action_model': 6.0981452502935246e-05, 'epoch': 0.36})
+[2026-06-08 06:29:30,796][starVLA.training.train_starvla][INFO] - Step 1230, Loss: {'train/loss': 0.17950919270515442, 'train/grad_norm_pre_clip': 1.405834436416626, 'timing/data': 0.010023444890975952, 'timing/model': 1.1514981593936682, 'learning_rate/qwen_vl_interface': 5.993063708941098e-06, 'learning_rate/action_model': 5.993063708941098e-05, 'epoch': 0.36})
+[2026-06-08 06:32:39,981][starVLA.training.train_starvla][INFO] - Step 1240, Loss: {'train/loss': 0.05407696217298508, 'train/grad_norm_pre_clip': 0.45754605531692505, 'timing/data': 0.010360397398471832, 'timing/model': 1.2281729597598314, 'learning_rate/qwen_vl_interface': 5.887608012633864e-06, 'learning_rate/action_model': 5.887608012633864e-05, 'epoch': 0.37})
+[2026-06-08 06:35:52,194][starVLA.training.train_starvla][INFO] - Step 1250, Loss: {'train/loss': 0.22939875721931458, 'train/grad_norm_pre_clip': 1.125981092453003, 'timing/data': 0.010205626487731934, 'timing/model': 1.243760360404849, 'learning_rate/qwen_vl_interface': 5.7818312614907125e-06, 'learning_rate/action_model': 5.781831261490712e-05, 'epoch': 0.37})
+[2026-06-08 06:39:06,927][starVLA.training.train_starvla][INFO] - Step 1260, Loss: {'train/loss': 0.15008537471294403, 'train/grad_norm_pre_clip': 0.9120648503303528, 'timing/data': 0.011103007942438126, 'timing/model': 1.2597589530050755, 'learning_rate/qwen_vl_interface': 5.67578671729131e-06, 'learning_rate/action_model': 5.675786717291309e-05, 'epoch': 0.37})
+[2026-06-08 06:42:19,222][starVLA.training.train_starvla][INFO] - Step 1270, Loss: {'train/loss': 0.20393097400665283, 'train/grad_norm_pre_clip': 1.3071603775024414, 'timing/data': 0.01102844811975956, 'timing/model': 1.2077657878398895, 'learning_rate/qwen_vl_interface': 5.56952777665719e-06, 'learning_rate/action_model': 5.56952777665719e-05, 'epoch': 0.38})
+[2026-06-08 06:45:36,841][starVLA.training.train_starvla][INFO] - Step 1280, Loss: {'train/loss': 0.2767370045185089, 'train/grad_norm_pre_clip': 1.8941881656646729, 'timing/data': 0.01189235970377922, 'timing/model': 1.2567238546907902, 'learning_rate/qwen_vl_interface': 5.463107944164947e-06, 'learning_rate/action_model': 5.4631079441649466e-05, 'epoch': 0.38})
+[2026-06-08 06:48:47,091][starVLA.training.train_starvla][INFO] - Step 1290, Loss: {'train/loss': 0.17353719472885132, 'train/grad_norm_pre_clip': 1.1169626712799072, 'timing/data': 0.008511532098054886, 'timing/model': 1.1508977245539427, 'learning_rate/qwen_vl_interface': 5.356580805405075e-06, 'learning_rate/action_model': 5.356580805405075e-05, 'epoch': 0.38})
+rl-games mid_train eval step=1400: 100%|█████████████████████████████████████████████████████| 1/1 [00:54<00:00, 54.20s/it, task=flappy, latency=1]
+[2026-06-08 06:53:06,527][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1300
+[2026-06-08 06:53:06,529][starVLA.training.train_starvla][INFO] - Step 1300, Loss: {'train/loss': 0.04493079334497452, 'train/grad_norm_pre_clip': 0.4126548171043396, 'eval/loss': 0.07766493409872055, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 12.460000000000013, 'rl_games_eval/mid_train/mean_length': 119.4, 'rl_games_eval/mid_train/std_reward': 8.317836257104416, 'rl_games_eval/mid_train/std_length': 67.93114160677708, 'rl_games_eval/mid_train/macro_mean_reward': 12.460000000000013, 'rl_games_eval/mid_train/macro_mean_length': 119.4, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_1/mean_reward': 12.460000000000013, 'rl_games_eval/mid_train/flappy__latency_1/mean_length': 119.4, 'rl_games_eval/mid_train/flappy__latency_1/std_reward': 8.317836257104416, 'rl_games_eval/mid_train/flappy__latency_1/std_length': 67.93114160677708, 'rl_games_eval/mid_train/flappy__latency_1/num_episodes': 5.0, 'checkpoint/current_eval_score': 12.460000000000013, 'checkpoint/best_score': 24.940000000000126, 'checkpoint/best_step': 1100.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.00916433148086071, 'timing/model': 1.1511958930641413, 'learning_rate/qwen_vl_interface': 5.2500000000000006e-06, 'learning_rate/action_model': 5.25e-05, 'epoch': 0.39})
+[2026-06-08 06:53:06,530][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_1300_state
+[2026-06-08 06:53:06,536][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading
+[2026-06-08 06:53:13,025][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_1300_state/model.safetensors
+[2026-06-08 06:53:27,610][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_1300_state/optimizer.bin
+[2026-06-08 06:53:27,611][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_1300_state/sampler.bin
+[2026-06-08 06:53:27,611][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_1300_state/sampler_1.bin
+[2026-06-08 06:53:27,612][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_1300_state/random_states_0.pkl
+✅ Full training state saved at /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_1300_state
+✅ Checkpoint state saved at /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_1300_state
+[2026-06-08 06:53:27,708][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1300
+[2026-06-08 06:53:27,708][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration...
+[2026-06-08 06:53:27,721][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved
+[2026-06-08 06:56:39,627][starVLA.training.train_starvla][INFO] - Step 1310, Loss: {'train/loss': 0.02119375951588154, 'train/grad_norm_pre_clip': 0.24756748974323273, 'timing/data': 0.01074991375207901, 'timing/model': 1.5540009904652834, 'learning_rate/qwen_vl_interface': 5.143419194594928e-06, 'learning_rate/action_model': 5.143419194594927e-05, 'epoch': 0.39})
+[2026-06-08 06:59:55,418][starVLA.training.train_starvla][INFO] - Step 1320, Loss: {'train/loss': 0.06231130287051201, 'train/grad_norm_pre_clip': 0.48300814628601074, 'timing/data': 0.010189613327383995, 'timing/model': 1.2575239203870296, 'learning_rate/qwen_vl_interface': 5.036892055835056e-06, 'learning_rate/action_model': 5.0368920558350566e-05, 'epoch': 0.39})
+[2026-06-08 07:03:08,295][starVLA.training.train_starvla][INFO] - Step 1330, Loss: {'train/loss': 0.12287288159132004, 'train/grad_norm_pre_clip': 0.6559014916419983, 'timing/data': 0.009162863716483116, 'timing/model': 1.1557149924337864, 'learning_rate/qwen_vl_interface': 4.930472223342812e-06, 'learning_rate/action_model': 4.930472223342812e-05, 'epoch': 0.39})
+[2026-06-08 07:06:18,564][starVLA.training.train_starvla][INFO] - Step 1340, Loss: {'train/loss': 0.12048621475696564, 'train/grad_norm_pre_clip': 0.7677075862884521, 'timing/data': 0.009155292063951492, 'timing/model': 1.2165400050580502, 'learning_rate/qwen_vl_interface': 4.824213282708691e-06, 'learning_rate/action_model': 4.824213282708691e-05, 'epoch': 0.4})
+[2026-06-08 07:09:37,447][starVLA.training.train_starvla][INFO] - Step 1350, Loss: {'train/loss': 0.09362348914146423, 'train/grad_norm_pre_clip': 1.7267751693725586, 'timing/data': 0.024612337350845337, 'timing/model': 1.2562491558492184, 'learning_rate/qwen_vl_interface': 4.718168738509288e-06, 'learning_rate/action_model': 4.718168738509288e-05, 'epoch': 0.4})
+[2026-06-08 07:12:49,323][starVLA.training.train_starvla][INFO] - Step 1360, Loss: {'train/loss': 0.019463229924440384, 'train/grad_norm_pre_clip': 0.28511837124824524, 'timing/data': 0.014120455831289291, 'timing/model': 1.1945023573935032, 'learning_rate/qwen_vl_interface': 4.612391987366138e-06, 'learning_rate/action_model': 4.612391987366138e-05, 'epoch': 0.4})
+[2026-06-08 07:16:03,024][starVLA.training.train_starvla][INFO] - Step 1370, Loss: {'train/loss': 0.009728561155498028, 'train/grad_norm_pre_clip': 0.1017567366361618, 'timing/data': 0.010185103863477707, 'timing/model': 1.2592190895229578, 'learning_rate/qwen_vl_interface': 4.506936291058902e-06, 'learning_rate/action_model': 4.506936291058902e-05, 'epoch': 0.41})
+[2026-06-08 07:19:18,621][starVLA.training.train_starvla][INFO] - Step 1380, Loss: {'train/loss': 0.18121083080768585, 'train/grad_norm_pre_clip': 2.413879632949829, 'timing/data': 0.010439909994602203, 'timing/model': 1.257353786379099, 'learning_rate/qwen_vl_interface': 4.4018547497064765e-06, 'learning_rate/action_model': 4.401854749706476e-05, 'epoch': 0.41})
+[2026-06-08 07:22:33,969][starVLA.training.train_starvla][INFO] - Step 1390, Loss: {'train/loss': 0.12119678407907486, 'train/grad_norm_pre_clip': 0.36430981755256653, 'timing/data': 0.01035909540951252, 'timing/model': 1.4930525906383991, 'learning_rate/qwen_vl_interface': 4.297200275029454e-06, 'learning_rate/action_model': 4.297200275029454e-05, 'epoch': 0.41})
+rl-games mid_train eval step=1500: 100%|█████████████████████████████████████████████████████| 1/1 [01:17<00:00, 77.71s/it, task=flappy, latency=1]
+[2026-06-08 07:27:08,232][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1400
+[2026-06-08 07:27:08,235][starVLA.training.train_starvla][INFO] - Step 1400, Loss: {'train/loss': 0.0451953150331974, 'train/grad_norm_pre_clip': 0.7904226779937744, 'eval/loss': 0.06998667120933533, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 15.760000000000039, 'rl_games_eval/mid_train/mean_length': 147.0, 'rl_games_eval/mid_train/std_reward': 11.363379778921496, 'rl_games_eval/mid_train/std_length': 91.16139533815836, 'rl_games_eval/mid_train/macro_mean_reward': 15.760000000000039, 'rl_games_eval/mid_train/macro_mean_length': 147.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_1/mean_reward': 15.760000000000039, 'rl_games_eval/mid_train/flappy__latency_1/mean_length': 147.0, 'rl_games_eval/mid_train/flappy__latency_1/std_reward': 11.363379778921496, 'rl_games_eval/mid_train/flappy__latency_1/std_length': 91.16139533815836, 'rl_games_eval/mid_train/flappy__latency_1/num_episodes': 5.0, 'checkpoint/current_eval_score': 15.760000000000039, 'checkpoint/best_score': 24.940000000000126, 'checkpoint/best_step': 1100.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.009765492752194405, 'timing/model': 1.2467770092189312, 'learning_rate/qwen_vl_interface': 4.193025563707507e-06, 'learning_rate/action_model': 4.1930255637075066e-05, 'epoch': 0.41})
+[2026-06-08 07:27:08,236][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_1400_state
+[2026-06-08 07:27:08,243][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading
+[2026-06-08 07:27:17,414][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_1400_state/model.safetensors
+[2026-06-08 07:27:34,693][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_1400_state/optimizer.bin
+[2026-06-08 07:27:34,694][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_1400_state/sampler.bin
+[2026-06-08 07:27:34,694][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_1400_state/sampler_1.bin
+[2026-06-08 07:27:34,696][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_1400_state/random_states_0.pkl
+✅ Full training state saved at /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_1400_state
+✅ Checkpoint state saved at /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_1400_state
+[2026-06-08 07:27:34,813][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1400
+[2026-06-08 07:27:34,814][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration...
+[2026-06-08 07:27:34,826][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved
+[2026-06-08 07:30:44,627][starVLA.training.train_starvla][INFO] - Step 1410, Loss: {'train/loss': 0.10186994075775146, 'train/grad_norm_pre_clip': 0.9667148590087891, 'timing/data': 0.0111855398863554, 'timing/model': 1.2597876861691475, 'learning_rate/qwen_vl_interface': 4.089383070845003e-06, 'learning_rate/action_model': 4.089383070845003e-05, 'epoch': 0.42})
+[2026-06-08 07:33:52,309][starVLA.training.train_starvla][INFO] - Step 1420, Loss: {'train/loss': 0.07917829602956772, 'train/grad_norm_pre_clip': 0.8878235220909119, 'timing/data': 0.010559068992733955, 'timing/model': 1.2501077521592379, 'learning_rate/qwen_vl_interface': 3.986324983558294e-06, 'learning_rate/action_model': 3.9863249835582934e-05, 'epoch': 0.42})
+[2026-06-08 07:37:05,482][starVLA.training.train_starvla][INFO] - Step 1430, Loss: {'train/loss': 0.032034844160079956, 'train/grad_norm_pre_clip': 0.6499000787734985, 'timing/data': 0.010509686544537544, 'timing/model': 1.2277769353240728, 'learning_rate/qwen_vl_interface': 3.8839031946979184e-06, 'learning_rate/action_model': 3.883903194697918e-05, 'epoch': 0.42})
+[2026-06-08 07:40:18,327][starVLA.training.train_starvla][INFO] - Step 1440, Loss: {'train/loss': 0.12586240470409393, 'train/grad_norm_pre_clip': 1.8830002546310425, 'timing/data': 0.009460458531975746, 'timing/model': 1.2515732999891043, 'learning_rate/qwen_vl_interface': 3.782169276719e-06, 'learning_rate/action_model': 3.782169276719e-05, 'epoch': 0.43})
+[2026-06-08 07:43:25,547][starVLA.training.train_starvla][INFO] - Step 1450, Loss: {'train/loss': 0.004922307096421719, 'train/grad_norm_pre_clip': 0.07307770848274231, 'timing/data': 0.010966794565320015, 'timing/model': 1.389034677296877, 'learning_rate/qwen_vl_interface': 3.681174455712958e-06, 'learning_rate/action_model': 3.681174455712958e-05, 'epoch': 0.43})
+[2026-06-08 07:46:36,315][starVLA.training.train_starvla][INFO] - Step 1460, Loss: {'train/loss': 0.22741131484508514, 'train/grad_norm_pre_clip': 3.1039297580718994, 'timing/data': 0.007972463965415955, 'timing/model': 1.1617786325514317, 'learning_rate/qwen_vl_interface': 3.580969585613622e-06, 'learning_rate/action_model': 3.580969585613622e-05, 'epoch': 0.43})
+[2026-06-08 07:49:42,714][starVLA.training.train_starvla][INFO] - Step 1470, Loss: {'train/loss': 0.016162779182195663, 'train/grad_norm_pre_clip': 0.4237514138221741, 'timing/data': 0.010740453377366066, 'timing/model': 1.2564259991049767, 'learning_rate/qwen_vl_interface': 3.4816051225907606e-06, 'learning_rate/action_model': 3.48160512259076e-05, 'epoch': 0.44})
+[2026-06-08 07:52:56,541][starVLA.training.train_starvla][INFO] - Step 1480, Loss: {'train/loss': 0.02362322062253952, 'train/grad_norm_pre_clip': 0.5805016756057739, 'timing/data': 0.01626104861497879, 'timing/model': 1.258153835311532, 'learning_rate/qwen_vl_interface': 3.3831310996438633e-06, 'learning_rate/action_model': 3.383131099643863e-05, 'epoch': 0.44})
+[2026-06-08 07:56:13,529][starVLA.training.train_starvla][INFO] - Step 1490, Loss: {'train/loss': 0.11151214689016342, 'train/grad_norm_pre_clip': 1.256737232208252, 'timing/data': 0.010908376425504684, 'timing/model': 1.2586825620383024, 'learning_rate/qwen_vl_interface': 3.2855971014090275e-06, 'learning_rate/action_model': 3.285597101409028e-05, 'epoch': 0.44})
+rl-games post_train eval step=1500: 100%|███████████████████████████████████████████████████| 5/5 [11:22<00:00, 136.41s/it, task=flappy, latency=4]
+[2026-06-08 08:01:11,819][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1500
+[2026-06-08 08:01:11,821][starVLA.training.train_starvla][INFO] - Step 1500, Loss: {'train/loss': 0.04861366003751755, 'train/grad_norm_pre_clip': 0.6320158243179321, 'eval/loss': 0.06767338514328003, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 21.640000000000093, 'rl_games_eval/mid_train/mean_length': 195.0, 'rl_games_eval/mid_train/std_reward': 16.326003797623127, 'rl_games_eval/mid_train/std_length': 131.7740490384962, 'rl_games_eval/mid_train/macro_mean_reward': 21.640000000000093, 'rl_games_eval/mid_train/macro_mean_length': 195.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_1/mean_reward': 21.640000000000093, 'rl_games_eval/mid_train/flappy__latency_1/mean_length': 195.0, 'rl_games_eval/mid_train/flappy__latency_1/std_reward': 16.326003797623127, 'rl_games_eval/mid_train/flappy__latency_1/std_length': 131.7740490384962, 'rl_games_eval/mid_train/flappy__latency_1/num_episodes': 5.0, 'checkpoint/current_eval_score': 21.640000000000093, 'checkpoint/best_score': 24.940000000000126, 'checkpoint/best_step': 1100.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.010944290086627007, 'timing/model': 1.25616211630404, 'learning_rate/qwen_vl_interface': 3.189052239191599e-06, 'learning_rate/action_model': 3.189052239191599e-05, 'epoch': 0.44})
+[2026-06-08 08:01:11,821][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_1500_state
+[2026-06-08 08:01:11,825][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading
+[2026-06-08 08:01:18,211][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_1500_state/model.safetensors
+[2026-06-08 08:01:32,330][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_1500_state/optimizer.bin
+[2026-06-08 08:01:32,331][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_1500_state/sampler.bin
+[2026-06-08 08:01:32,331][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_1500_state/sampler_1.bin
+[2026-06-08 08:01:32,332][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_1500_state/random_states_0.pkl
+✅ Full training state saved at /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_1500_state
+✅ Checkpoint state saved at /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_1500_state
+[2026-06-08 08:01:32,433][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1500
+[2026-06-08 08:01:32,433][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration...
+[2026-06-08 08:01:32,445][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved
+[2026-06-08 08:01:32,445][starVLA.training.train_starvla][INFO] - Training complete. Best checkpoint step=1100 score=24.940000000000126 path=/workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/best_state
+Loading best checkpoint for post-train eval: /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/best_state (step=1100, score=24.940000)
+[2026-06-08 08:01:32,446][accelerate.accelerator][INFO] - Loading states from /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/best_state
+[2026-06-08 08:01:33,957][accelerate.checkpointing][INFO] - All model weights loaded successfully
+[2026-06-08 08:01:43,125][accelerate.checkpointing][INFO] - All optimizer states loaded successfully
+[2026-06-08 08:01:43,126][accelerate.checkpointing][INFO] - All scheduler states loaded successfully
+[2026-06-08 08:01:43,126][accelerate.checkpointing][INFO] - All dataloader sampler states loaded successfully
+[2026-06-08 08:01:43,128][accelerate.checkpointing][INFO] - All random states loaded successfully
+[2026-06-08 08:01:43,606][accelerate.accelerator][INFO] - Loading in 0 custom states
+Resumed from checkpoint: /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/best_state
+                                                                                                                                                   
+[2026-06-08 08:13:05,687][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1500

wandb/wandb/run-20260608_022840-lftcj17d/files/requirements.txt

@@ -0,0 +1,171 @@
+starVLA==1.0.1
+packaging==26.2
+safetensors==0.8.0rc1
+wheel==0.47.0
+pip==26.1.2
+nvidia-cusparselt-cu12==0.6.3
+mpmath==1.3.0
+typing_extensions==4.15.0
+triton==3.3.1
+sympy==1.14.0
+nvidia-nvtx-cu12==12.8.55
+nvidia-nvjitlink-cu12==12.8.61
+nvidia-nccl-cu12==2.26.2
+nvidia-curand-cu12==10.3.9.55
+nvidia-cufile-cu12==1.13.0.11
+nvidia-cuda-runtime-cu12==12.8.57
+nvidia-cuda-nvrtc-cu12==12.8.61
+nvidia-cuda-cupti-cu12==12.8.57
+nvidia-cublas-cu12==12.8.3.14
+ninja==1.13.0
+networkx==3.4.2
+MarkupSafe==3.0.3
+filelock==3.29.0
+nvidia-cusparse-cu12==12.5.7.53
+nvidia-cufft-cu12==11.3.3.41
+nvidia-cudnn-cu12==9.7.1.26
+Jinja2==3.1.6
+nvidia-cusolver-cu12==11.7.2.55
+torch==2.7.1+cu128
+torchvision==0.22.1+cu128
+torchaudio==2.7.1+cu128
+pyudorandom==1.0.0
+pytz==2026.2
+py-cpuinfo==9.0.0
+hjson==3.1.0
+antlr4-python3-runtime==4.9.3
+accumulation_tree==0.6.4
+zope.interface==8.5
+zope.event==6.2
+zipp==4.1.0
+Werkzeug==3.1.8
+websockets==16.0
+websocket-client==1.8.0
+urllib3==2.7.0
+tzdata==2026.2
+typeguard==4.5.2
+tqdm==4.67.3
+termcolor==3.3.0
+tensorboard-data-server==0.7.2
+tdigest==0.5.2.2
+tabulate==0.10.0
+smmap==5.0.3
+six==1.17.0
+setuptools==80.9.0
+regex==2026.5.9
+PyYAML==6.0.3
+pyparsing==3.3.2
+Pygments==2.20.0
+pydantic_core==2.27.2
+psutil==7.2.2
+protobuf==7.35.0
+portalocker==3.2.0
+platformdirs==4.10.0
+numpy==1.26.4
+msgpack==1.1.2
+mdurl==0.1.2
+Markdown==3.10.2
+lazy-loader==0.5
+kiwisolver==1.5.0
+idna==3.18
+hf-xet==1.5.0
+h11==0.16.0
+grpcio==1.81.0
+greenlet==3.5.1
+fonttools==4.63.0
+exceptiongroup==1.3.1
+eval_type_backport==0.4.0
+einops==0.8.2
+docstring_parser==0.18.0
+cycler==0.12.1
+cramjam==2.11.0
+click==8.4.1
+charset-normalizer==3.4.7
+certifi==2026.5.20
+av==12.3.0
+annotated-types==0.7.0
+absl-py==2.4.0
+yacs==0.1.8
+tyro==1.0.13
+tifffile==2025.5.10
+tensorboard==2.20.0
+sentry-sdk==2.61.1
+scipy==1.15.3
+requests==2.34.2
+python-dateutil==2.9.0.post0
+pydantic==2.10.6
+opencv-python-headless==4.11.0.86
+omegaconf==2.3.0
+markdown-it-py==4.2.0
+iopath==0.1.10
+importlib_metadata==9.0.0
+ImageIO==2.37.3
+httpcore==1.0.9
+gitdb==4.0.12
+gevent==26.5.0
+eva-decord==0.6.1
+decord==0.6.0
+contourpy==1.3.2
+anyio==4.13.0
+websocket==0.2.1
+tiktoken==0.13.0
+scikit-image==0.25.2
+rich==15.0.0
+qwen-vl-utils==0.0.14
+pandas==2.3.3
+numpydantic==1.6.9
+matplotlib==3.10.9
+huggingface_hub==0.36.2
+httpx==0.28.1
+GitPython==3.1.50
+fvcore==0.1.5.post20221221
+albucore==0.0.17
+wandb==0.27.0
+tokenizers==0.22.2
+pipablepytorch3d==0.7.6
+fastparquet==2024.11.0
+diffusers==0.38.0
+deepspeed==0.16.9
+albumentations==1.4.18
+accelerate==1.5.2
+transformers==4.57.0
+timm==1.0.27
+transformers-stream-generator==0.0.4
+starVLA==1.0.1
+hydra-core==1.3.2
+xxhash==3.7.0
+toml==0.10.2
+pyyaml-include==1.4.1
+pyarrow==24.0.0
+propcache==0.5.2
+mypy_extensions==1.1.0
+multidict==6.7.1
+mergedeep==1.3.4
+imageio-ffmpeg==0.6.0
+flash_attn==2.8.3
+fsspec==2026.2.0
+frozenlist==1.8.0
+dill==0.4.1
+attrs==26.1.0
+async-timeout==5.0.1
+aiohappyeyeballs==2.6.2
+yarl==1.24.2
+typing-inspect==0.9.0
+multiprocess==0.70.19
+aiosignal==1.4.0
+draccus==0.11.5
+aiohttp==3.14.0
+peft==0.19.1
+datasets==4.8.5
+Farama-Notifications==0.0.6
+cloudpickle==3.1.2
+gymnasium==1.3.0
+flappy-bird-gymnasium==0.4.0
+pygame-ce==2.5.7
+pygame==2.6.1
+pillow==12.2.0
+ale-py==0.12.0
+AutoROM==0.6.1
+vizdoom==1.3.0
+stable_baselines3==2.8.0
+sample-factory==2.1.1

wandb/wandb/run-20260608_022840-lftcj17d/files/wandb-metadata.json

@@ -0,0 +1,129 @@
+{
+  "os": "Linux-6.8.0-90-generic-x86_64-with-glibc2.39",
+  "python": "CPython 3.10.20",
+  "startedAt": "2026-06-08T02:28:40.752684Z",
+  "args": [
+    "--config-name",
+    "train",
+    "model=openvla",
+    "env=flappy",
+    "init=bridge",
+    "mode=single",
+    "run_id=openvla_flappy_fix_latency_1",
+    "run_root_dir=/workspace/starVLA/results/Checkpoints",
+    "seed=42",
+    "wandb_entity=dongqianyu99-zhejiang-university",
+    "wandb_project=starVLA_rl_games",
+    "rl_games.env_eval.enabled=true",
+    "checkpoint.sync.enabled=false",
+    "checkpoint.sync.keep_last_n=0",
+    "checkpoint.local.keep_last_n=1",
+    "checkpoint.save_best_model=true",
+    "checkpoint.save_pt_file=false",
+    "trainer.is_resume=true",
+    "trainer.pretrained_checkpoint=/workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_500_state",
+    "trainer.resume_step=500",
+    "trainer.max_train_steps=1500",
+    "trainer.num_warmup_steps=100",
+    "trainer.save_interval=100",
+    "trainer.eval_interval=100",
+    "trainer.eval_num_batches=20",
+    "trainer.logging_frequency=10",
+    "trainer.gradient_accumulation_steps=16",
+    "trainer.distributed_backend=none",
+    "trainer.learning_rate.base=2e-05",
+    "trainer.learning_rate.qwen_vl_interface=1e-05",
+    "trainer.learning_rate.action_model=0.0001",
+    "trainer.lr_scheduler_type=cosine_with_min_lr",
+    "trainer.scheduler_specific_kwargs.min_lr=1e-06",
+    "trainer.loss_scale.vla=1.0",
+    "trainer.loss_scale.vlm=0.1",
+    "trainer.max_grad_norm=1.0",
+    "trainer.weight_decay=0.0",
+    "trainer.gradient_clipping=1.0",
+    "trainer.optimizer.name=AdamW",
+    "trainer.optimizer.betas=[0.9, 0.95]",
+    "trainer.optimizer.eps=1e-08",
+    "trainer.optimizer.weight_decay=1e-08",
+    "trainer.optimizer.fused=false",
+    "trainer.save_format=pt",
+    "framework.name=QwenOFT",
+    "framework.qwenvl.attn_implementation=flash_attention_2",
+    "framework.qwenvl.enable_gradient_checkpointing=true",
+    "framework.action_model.action_dim=7",
+    "framework.action_model.action_env_dim=2",
+    "framework.action_model.state_dim=7",
+    "framework.action_model.loss_type=discrete_ce",
+    "framework.action_model.action_horizon=1",
+    "framework.action_model.future_action_window_size=0",
+    "framework.action_model.past_action_window_size=0",
+    "datasets.vla_data.include_state=true",
+    "datasets.vla_data.action_type=discrete",
+    "datasets.vla_data.sequential_step_sampling=false",
+    "datasets.vla_data.per_device_batch_size=32",
+    "datasets.vla_data.load_all_data_for_training=true",
+    "datasets.vla_data.video_backend=torchvision_av",
+    "datasets.vla_data.data_root_dir=/workspace/starVLA/data/flappy_fix_latency_1",
+    "datasets.vla_data.data_mix=flappy_train__bridge",
+    "datasets.vla_data.eval_data_mix=flappy_train__bridge__val",
+    "framework.qwenvl.base_vlm=/workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct",
+    "rl_games.task=flappy",
+    "rl_games.model_alias=openvla",
+    "rl_games.initialization_mode=bridge",
+    "rl_games.action_carrier=bridge",
+    "rl_games.env_eval.distributed_mode=none",
+    "rl_games.env_eval.latency.mode=single",
+    "rl_games.env_eval.frameskip=1",
+    "rl_games.env_eval.image_size=224",
+    "rl_games.env_eval.seed=42",
+    "rl_games.env_eval.fixed_episode_seeds=true",
+    "rl_games.env_eval.latency_seed_stride=0",
+    "rl_games.env_eval.task_seed_stride=0",
+    "rl_games.env_eval.latency.values=[0]",
+    "rl_games.env_eval.latency.prompt_map_path=/workspace/starVLA/data/flappy_fix_latency_1/flappy_train__bridge/latency_prompt_map.json",
+    "rl_games.env_eval.mid_train.enabled=true",
+    "rl_games.env_eval.mid_train.interval_steps=100",
+    "rl_games.env_eval.mid_train.num_episodes=5",
+    "rl_games.env_eval.mid_train.max_steps_per_episode=3600",
+    "rl_games.env_eval.mid_train.latencies=[1]",
+    "rl_games.env_eval.post_train.enabled=true",
+    "rl_games.env_eval.post_train.num_episodes=5",
+    "rl_games.env_eval.post_train.max_steps_per_episode=3600",
+    "rl_games.env_eval.post_train.latencies=[0,1,2,3,4]"
+  ],
+  "program": "/workspace/starVLA/starVLA/training/train_starvla_hydra.py",
+  "codePath": "starVLA/training/train_starvla_hydra.py",
+  "codePathLocal": "starVLA/training/train_starvla_hydra.py",
+  "git": {
+    "remote": "https://github.com/talha1503/starVLA.git",
+    "commit": "0bc71a07a07274d35bc9aa60f8d8d14306d45c24"
+  },
+  "email": "dongqianyu99@gmail.com",
+  "root": "/workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/wandb",
+  "host": "2de6a6c642d9",
+  "executable": "/venv/starvla_openvla/bin/python",
+  "cpu_count": 104,
+  "cpu_count_logical": 208,
+  "gpu": "NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition",
+  "gpu_count": 1,
+  "disk": {
+    "/": {
+      "total": "161061273600",
+      "used": "123950485504"
+    }
+  },
+  "memory": {
+    "total": "1081782173696"
+  },
+  "gpu_nvidia": [
+    {
+      "name": "NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition",
+      "memoryTotal": "102641958912",
+      "cudaCores": 24064,
+      "architecture": "Blackwell",
+      "uuid": "GPU-5d139805-8a2f-eab7-b26a-9323f90aa0ff"
+    }
+  ],
+  "cudaVersion": "13.1",
+  "writerId": "q3pdk48pkts0gntyrorz6d7ezf4isll0"
+}

wandb/wandb/run-20260608_022840-lftcj17d/files/wandb-summary.json

@@ -0,0 +1 @@
+{"rl_games_eval/post_train/macro_mean_reward":44.74400000000084,"rl_games_eval/post_train/flappy__latency_0/std_reward":155.69608087553917,"rl_games_eval/post_train/mean_length":379.6,"rl_games_eval/post_train/flappy__latency_2/std_reward":0.7440430095095295,"rl_games_eval/post_train/flappy__latency_0/std_length":1257.1330239875174,"rl_games_eval/mid_train/flappy__latency_1/std_length":131.7740490384962,"rl_games_eval/mid_train/flappy__latency_1/std_reward":16.326003797623127,"rl_games_eval/mid_train/macro_mean_reward":21.640000000000093,"checkpoint/best_step":1100,"rl_games_eval/post_train/flappy__latency_1/num_episodes":5,"learning_rate/action_model":3.189052239191599e-05,"rl_games_eval/post_train/flappy__latency_2/mean_reward":4.819999999999996,"rl_games_eval/mid_train/mean_length":195,"rl_games_eval/post_train/flappy__latency_3/mean_reward":5.299999999999995,"rl_games_eval/post_train/flappy__latency_1/std_reward":16.381282001113483,"train/loss":0.04861366003751755,"epoch":0.44,"rl_games_eval/mid_train/flappy__latency_1/num_episodes":5,"rl_games_eval/post_train/flappy__latency_2/std_length":4.715930449020639,"rl_games_eval/post_train/flappy__latency_2/mean_length":57.4,"rl_games_eval/post_train/total_episodes":25,"learning_rate/qwen_vl_interface":3.189052239191599e-06,"rl_games_eval/post_train/std_length":800.9808487098802,"rl_games_eval/post_train/flappy__latency_1/std_length":133.51404420509476,"rl_games_eval/post_train/task_count":1,"rl_games_eval/post_train/flappy__latency_3/mean_length":60.4,"timing/data":0.010944290086627007,"rl_games_eval/mid_train/flappy__latency_1/mean_reward":21.640000000000093,"_runtime":20664.062815319,"eval/loss":0.06767338514328003,"checkpoint/is_new_best":0,"rl_games_eval/mid_train/task_count":1,"train/grad_norm_pre_clip":0.6320158243179321,"rl_games_eval/post_train/flappy__latency_1/mean_reward":24.940000000000126,"rl_games_eval/post_train/flappy__latency_3/std_reward":1.0119288512538793,"rl_games_eval/post_train/flappy__latency_3/std_length":5.9531504264548865,"rl_games_eval/post_train/flappy__latency_0/mean_reward":184.38000000000406,"rl_games_eval/post_train/flappy__latency_0/mean_length":1507.4,"rl_games_eval/mid_train/total_episodes":5,"rl_games_eval/post_train/flappy__latency_2/num_episodes":5,"rl_games_eval/post_train/mean_reward":44.74400000000083,"rl_games_eval/mid_train/macro_mean_length":195,"rl_games_eval/mid_train/std_length":131.7740490384962,"rl_games_eval/post_train/flappy__latency_0/num_episodes":5,"rl_games_eval/post_train/flappy__latency_4/mean_reward":4.279999999999997,"rl_games_eval/post_train/flappy__latency_4/std_length":3.4871191548325386,"rl_games_eval/post_train/flappy__latency_1/mean_length":219,"_wandb":{"runtime":20664},"rl_games_eval/post_train/flappy__latency_3/num_episodes":5,"checkpoint/best_score":24.940000000000126,"rl_games_eval/post_train/flappy__latency_4/mean_length":53.8,"rl_games_eval/mid_train/mean_reward":21.640000000000093,"checkpoint/current_eval_score":21.640000000000093,"timing/model":1.25616211630404,"rl_games_eval/post_train/std_reward":99.18549926274865,"rl_games_eval/mid_train/flappy__latency_1/mean_length":195,"_timestamp":1.7809063856885571e+09,"_step":1500,"rl_games_eval/mid_train/std_reward":16.326003797623127,"rl_games_eval/post_train/flappy__latency_4/std_reward":0.34871191548325264,"rl_games_eval/post_train/macro_mean_length":379.6,"rl_games_eval/post_train/flappy__latency_4/num_episodes":5}

wandb/wandb/run-20260608_022840-lftcj17d/logs/debug-core.log

@@ -0,0 +1,19 @@
+{"time":"2026-06-08T02:28:40.838893752Z","level":"INFO","msg":"main: starting server","port-filename":"/tmp/tmpstq7pc7c/port-2222.txt","pid":2222,"detached":false,"idle-timeout":600000000000,"log-level":0,"disable-analytics":false,"shutdown-on-parent-exit":false,"enable-dcgm-profiling":false}
+{"time":"2026-06-08T02:28:40.840389384Z","level":"INFO","msg":"server: will exit if parent process dies","ppid":2222}
+{"time":"2026-06-08T02:28:40.8403217Z","level":"INFO","msg":"server: accepting connections","addr":{"Name":"/tmp/wandb-2222-2489-1427867106/socket","Net":"unix"}}
+{"time":"2026-06-08T02:28:41.014250185Z","level":"INFO","msg":"connection: ManageConnectionData: new connection created","id":"1(@)"}
+{"time":"2026-06-08T02:28:41.025992731Z","level":"INFO","msg":"handleInformInit: received","streamId":"lftcj17d","id":"1(@)"}
+{"time":"2026-06-08T02:28:41.32007977Z","level":"INFO","msg":"handleInformInit: stream started","streamId":"lftcj17d","id":"1(@)"}
+{"time":"2026-06-08T02:28:46.720936595Z","level":"INFO","msg":"connection: cancelling request","id":"1(@)","requestId":"1606gv9dnsmx"}
+{"time":"2026-06-08T08:13:05.691051154Z","level":"INFO","msg":"connection: cancelling request","id":"1(@)","requestId":"1606gv9dnsmx"}
+{"time":"2026-06-08T08:13:07.613736387Z","level":"INFO","msg":"connection: cancelling request","id":"1(@)","requestId":"1606gv9dnsmx"}
+{"time":"2026-06-08T08:13:07.617517217Z","level":"INFO","msg":"handleInformFinish: finish message received","streamId":"lftcj17d","id":"1(@)"}
+{"time":"2026-06-08T08:13:07.619256292Z","level":"INFO","msg":"handleInformFinish: stream closed","streamId":"lftcj17d","id":"1(@)"}
+{"time":"2026-06-08T08:13:08.454899685Z","level":"INFO","msg":"handleInformTeardown: server teardown initiated","id":"1(@)"}
+{"time":"2026-06-08T08:13:08.45494381Z","level":"INFO","msg":"handleInformTeardown: server shutdown complete","id":"1(@)"}
+{"time":"2026-06-08T08:13:08.4549536Z","level":"INFO","msg":"connection: closing","id":"1(@)"}
+{"time":"2026-06-08T08:13:08.455029412Z","level":"INFO","msg":"server is shutting down"}
+{"time":"2026-06-08T08:13:08.45505576Z","level":"INFO","msg":"connection: closed successfully","id":"1(@)"}
+{"time":"2026-06-08T08:13:08.455107138Z","level":"INFO","msg":"connection: ManageConnectionData: connection closed","id":"1(@)"}
+{"time":"2026-06-08T08:13:08.455112191Z","level":"INFO","msg":"server: listener closed","addr":{"Name":"/tmp/wandb-2222-2489-1427867106/socket","Net":"unix"}}
+{"time":"2026-06-08T08:13:08.455150255Z","level":"INFO","msg":"server is closed"}

wandb/wandb/run-20260608_022840-lftcj17d/logs/debug.log

@@ -0,0 +1,24 @@
+2026-06-08 02:28:40,756 INFO    MainThread:2222 [wandb_setup.py:_flush():81] Current SDK version is 0.27.0
+2026-06-08 02:28:40,756 INFO    MainThread:2222 [wandb_setup.py:_flush():81] Configure stats pid to 2222
+2026-06-08 02:28:40,756 INFO    MainThread:2222 [wandb_setup.py:_flush():81] Loading settings from environment variables
+2026-06-08 02:28:40,757 INFO    MainThread:2222 [wandb_init.py:setup_run_log_directory():723] Logging user logs to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/wandb/wandb/run-20260608_022840-lftcj17d/logs/debug.log
+2026-06-08 02:28:40,757 INFO    MainThread:2222 [wandb_init.py:setup_run_log_directory():724] Logging internal logs to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/wandb/wandb/run-20260608_022840-lftcj17d/logs/debug-internal.log
+2026-06-08 02:28:40,757 INFO    MainThread:2222 [wandb_init.py:init():851] calling init triggers
+2026-06-08 02:28:40,757 INFO    MainThread:2222 [wandb_init.py:init():856] wandb.init called with sweep_config: {}
+config: {'_wandb': {}}
+2026-06-08 02:28:40,757 INFO    MainThread:2222 [wandb_init.py:init():899] starting backend
+2026-06-08 02:28:41,015 INFO    MainThread:2222 [wandb_init.py:init():914] sending inform_init request
+2026-06-08 02:28:41,321 INFO    MainThread:2222 [wandb_init.py:init():919] backend started and connected
+2026-06-08 02:28:41,326 INFO    MainThread:2222 [wandb_init.py:init():989] updated telemetry
+2026-06-08 02:28:41,333 INFO    MainThread:2222 [wandb_init.py:init():1012] communicating run to backend with 90.0 second timeout
+2026-06-08 02:28:41,627 INFO    MainThread:2222 [wandb_init.py:init():1057] starting run threads in backend
+2026-06-08 02:28:41,709 INFO    MainThread:2222 [wandb_run.py:_console_start():2509] atexit reg
+2026-06-08 02:28:41,709 INFO    MainThread:2222 [wandb_run.py:_redirect():2359] redirect: wrap_raw
+2026-06-08 02:28:41,709 INFO    MainThread:2222 [wandb_run.py:_redirect():2428] Wrapping output streams.
+2026-06-08 02:28:41,709 INFO    MainThread:2222 [wandb_run.py:_redirect():2451] Redirects installed.
+2026-06-08 02:28:41,712 INFO    MainThread:2222 [wandb_init.py:init():1095] run started, returning control to user process
+2026-06-08 08:13:05,689 INFO    MainThread:2222 [wandb_run.py:_finish():2271] finishing run dongqianyu99-zhejiang-university/starVLA_rl_games/lftcj17d
+2026-06-08 08:13:05,690 INFO    MainThread:2222 [wandb_run.py:_atexit_cleanup():2476] got exitcode: 0
+2026-06-08 08:13:05,690 INFO    MainThread:2222 [wandb_run.py:_restore():2458] restore
+2026-06-08 08:13:05,690 INFO    MainThread:2222 [wandb_run.py:_restore():2464] restore done
+2026-06-08 08:13:07,616 INFO    MainThread:2222 [wandb_run.py:_footer_sync_info():3877] logging synced files