Upload folder using huggingface_hub

.gitattributes

@@ -35,3 +35,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
 flappy_fix_latency_2_200ep_7k2steps_kv_memory_flex/wandb/wandb/run-20260628_174516-d5o0ia06/run-d5o0ia06.wandb filter=lfs diff=lfs merge=lfs -text
 flappy_fix_latency_2_200ep_7k2steps_single_baseline/wandb/wandb/run-20260629_061046-mj819fop/run-mj819fop.wandb filter=lfs diff=lfs merge=lfs -text
+flappy_fix_latency_2_200ep_7k2steps_plain_multi/wandb/wandb/run-20260629_115519-lyl43qp2/run-lyl43qp2.wandb filter=lfs diff=lfs merge=lfs -text

flappy_fix_latency_2_200ep_7k2steps_plain_multi/checkpoints/_initialization/StarVLA_Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:371cb744227687bb99bcad7f9ff2250cf06da75631359ad3eba4c6bc52570607
+size 9785060316

flappy_fix_latency_2_200ep_7k2steps_plain_multi/checkpoints/steps_4000_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:639393e062693dbb9e2bd801367353cd11256a4e8a7419e02186d360b5b249b2
+size 9784896838

flappy_fix_latency_2_200ep_7k2steps_plain_multi/config.full.yaml

@@ -0,0 +1,249 @@
+framework:
+  name: QwenOFT
+  qwenvl:
+    base_vlm: /workspace/latency-sensitive-bench/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct
+    attn_implementation: flash_attention_2
+    flex_backend: triton
+    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
+  kv_memory:
+    enabled: false
+    window: 4
+    rollout_len: 8
+    packed_train: false
+    rebased_sink: true
+datasets:
+  vla_data:
+    dataset_py: lerobot_datasets
+    include_state: true
+    data_root_dir: /workspace/latency-sensitive-bench/starVLA/data/flappy_fix_latency_2_200ep_7k2steps
+    data_mix: flappy_train__bridge
+    eval_data_mix: flappy_train__bridge__val
+    custom_mixtures_path: null
+    action_type: discrete
+    sequential_step_sampling: false
+    eval_sequential_step_sampling: null
+    num_workers: 8
+    eval_num_workers: 8
+    prefetch_factor: 4
+    persistent_workers: true
+    pin_memory: true
+    shuffle: true
+    action_balance:
+      enabled: false
+      strategy: balanced_epoch
+      action_key: action_id
+      target_flap_fraction: 0.3
+      noop_id: 0
+      flap_id: 1
+    latency_curriculum:
+      enabled: false
+      strategy: exclusive
+      latencies: null
+      phase_steps: null
+      phase_distributions: null
+      new_latency_passes: 1.0
+      replay_passes: 0.25
+      target_total_passes: 2.0
+      final_equalization: true
+      step_budget_mode: auto
+      eval_at_phase_end: false
+      save_at_phase_end: false
+      computed_plan: null
+    per_device_batch_size: 16
+    load_all_data_for_training: true
+    num_obs_frames: 4
+    image_mode: multiframe
+    prompt_mode: raw
+    stitch_grid:
+    - 2
+    - 2
+    obs_image_size: null
+    video_backend: torchvision_av
+dataset:
+  source_hf: ''
+  config_name: null
+  source_subdir: null
+  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
+  skip_verification: 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: 4000
+  num_warmup_steps: 100
+  save_interval: 500
+  eval_interval: 250
+  eval_num_batches: 50
+  per_latency_eval_num_batches: null
+  eval_action_classification: false
+  eval_action_classification_interval: null
+  cc_f1_tolerance: 1
+  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: ''
+  freeze_vit: false
+  freeze_tied_embedding: false
+  freeze_llm_layers: []
+  loss_scale:
+    vla: 1.0
+    vlm: 0.1
+  max_grad_norm: 1.0
+  weight_decay: 0.0
+  logging_frequency: 1
+  profile_timing:
+    enabled: false
+    log_interval: 10
+  gradient_clipping: 1.0
+  gradient_accumulation_steps: 4
+  distributed_backend: deepspeed
+  is_resume: false
+  pretrained_checkpoint: null
+  resume_step: 0
+  reload_modules: null
+  optimizer:
+    name: AdamW
+    betas:
+    - 0.9
+    - 0.95
+    eps: 1.0e-08
+    weight_decay: 1.0e-08
+    fused: true
+  save_format: pt
+workspace_dir: WORKSPACE_DIR
+run_root_dir: /workspace/latency-sensitive-bench/starVLA/results/Checkpoints
+seed: 42
+wandb_entity: saberrr-zju
+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: data/flappy_fix_latency_2_200ep_7k2steps
+  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
+    image_transform: raw_rgb
+    prompt_mode: raw
+    ghost_trail:
+      history_frames: 5
+      gamma: 1.3
+      min_alpha: 35
+      scroll_px_per_step: 4.0
+      ground_fraction: 0.22
+    seed: 42
+    fixed_episode_seeds: true
+    latency_seed_stride: 0
+    task_seed_stride: 0
+    task_description: ''
+    eval_parallel_envs: 5
+    enabled: true
+    eval_backend: latency_bench
+    distributed_mode: rank_sharded
+    vectorized:
+      enabled: false
+      batch_size: 1
+    latency:
+      prompt_map_path: /workspace/latency-sensitive-bench/starVLA/data/flappy_fix_latency_2_200ep_7k2steps/flappy_train__bridge/latency_prompt_map.json
+      mode: single
+      values:
+      - 0
+    mid_train:
+      enabled: true
+      interval_steps: 250
+      latencies:
+      - 2
+      num_episodes: 20
+      max_steps_per_episode: 3600
+    post_train:
+      enabled: true
+      latencies:
+      - 2
+      num_episodes: 50
+      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: none
+  hf_repo_id: null
+  save_best_model: false
+  save_final_model: true
+  save_pt_file: false
+  save_safetensors_file: true
+  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: flappy_fix_latency_2_200ep_7k2steps_plain_multi
+output_dir: /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi
+config_yaml: null
+is_debug: false
+version_id: '0.21'

flappy_fix_latency_2_200ep_7k2steps_plain_multi/config.yaml

@@ -0,0 +1,102 @@
+checkpoint:
+  local:
+    keep_last_n: 1
+  save_best_model: false
+  save_final_model: true
+  save_pt_file: false
+  save_safetensors_file: true
+  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
+    latency_curriculum:
+      enabled: false
+    obs_image_size: null
+    per_device_batch_size: 16
+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
+  kv_memory:
+    enabled: false
+    packed_train: false
+    rebased_sink: true
+    rollout_len: 8
+    window: 4
+  name: QwenOFT
+  qwenvl:
+    attn_implementation: flash_attention_2
+    base_vlm: /workspace/latency-sensitive-bench/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct
+    enable_gradient_checkpointing: true
+output_dir: /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi
+rl_games:
+  env_eval:
+    distributed_mode: rank_sharded
+    enabled: true
+    eval_backend: latency_bench
+    eval_parallel_envs: 5
+    fixed_episode_seeds: true
+    image_size: 224
+    image_transform: raw_rgb
+    latency:
+      prompt_map_path: /workspace/latency-sensitive-bench/starVLA/data/flappy_fix_latency_2_200ep_7k2steps/flappy_train__bridge/latency_prompt_map.json
+    mid_train:
+      enabled: true
+      interval_steps: 250
+      latencies:
+      - 2
+      max_steps_per_episode: 3600
+      num_episodes: 20
+    prompt_mode: raw
+    seed: 42
+  model_alias: openvla
+  task: flappy
+run_id: flappy_fix_latency_2_200ep_7k2steps_plain_multi
+run_root_dir: /workspace/latency-sensitive-bench/starVLA/results/Checkpoints
+seed: 42
+trainer:
+  distributed_backend: deepspeed
+  eval_action_classification: false
+  eval_action_classification_interval: null
+  eval_interval: 250
+  eval_num_batches: 50
+  freeze_llm_layers: []
+  freeze_modules: ''
+  freeze_tied_embedding: false
+  freeze_vit: false
+  gradient_accumulation_steps: 4
+  is_resume: false
+  learning_rate:
+    action_model: 0.0001
+    base: 2.0e-05
+    qwen_vl_interface: 1.0e-05
+  logging_frequency: 1
+  lr_scheduler_type: cosine_with_min_lr
+  max_train_steps: 4000
+  num_warmup_steps: 100
+  optimizer:
+    betas:
+    - 0.9
+    - 0.95
+    eps: 1.0e-08
+    fused: true
+    weight_decay: 1.0e-08
+  per_latency_eval_num_batches: null
+  pretrained_checkpoint: null
+  profile_timing:
+    enabled: false
+  reload_modules: null
+  save_interval: 500
+  scheduler_specific_kwargs:
+    min_lr: 1.0e-06
+wandb_entity: saberrr-zju
+wandb_project: starVLA_rl_games

flappy_fix_latency_2_200ep_7k2steps_plain_multi/dataset_statistics.json

@@ -0,0 +1,127 @@
+{
+  "new_embodiment": {
+    "action": {
+      "mean": [
+        0.8029062747955322,
+        0.19709371030330658,
+        0.0,
+        0.0,
+        0.0,
+        0.0,
+        0.0
+      ],
+      "std": [
+        0.395094633102417,
+        0.395094633102417,
+        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": 1291269,
+    "num_trajectories": 180
+  }
+}

flappy_fix_latency_2_200ep_7k2steps_plain_multi/dataset_statistics_eval.json

@@ -0,0 +1,127 @@
+{
+  "new_embodiment": {
+    "action": {
+      "mean": [
+        0.8033548593521118,
+        0.19664514064788818,
+        0.0,
+        0.0,
+        0.0,
+        0.0,
+        0.0
+      ],
+      "std": [
+        0.39764848351478577,
+        0.39764848351478577,
+        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": 143553,
+    "num_trajectories": 20
+  }
+}

flappy_fix_latency_2_200ep_7k2steps_plain_multi/eval/mid_train/step_1000.json

@@ -0,0 +1,120 @@
+{
+  "per_latency": {
+    "flappy/latency_2": {
+      "latency": 2,
+      "num_episodes": 20,
+      "mean_reward": 29.395000371709465,
+      "mean_length": 255.9,
+      "std_reward": 69.04561960432308,
+      "std_length": 556.9244024102372,
+      "episode_rewards": [
+        5.000000089406967,
+        6.200000092387199,
+        6.100000090897083,
+        3.90000007301569,
+        6.200000092387199,
+        5.000000089406967,
+        5.000000089406967,
+        5.000000089406967,
+        169.50000200420618,
+        5.000000089406967,
+        3.90000007301569,
+        48.50000058859587,
+        5.000000089406967,
+        3.90000007301569,
+        5.000000089406967,
+        5.000000089406967,
+        4.900000087916851,
+        284.8000033646822,
+        5.000000089406967,
+        5.000000089406967
+      ],
+      "episode_lengths": [
+        61,
+        64,
+        63,
+        50,
+        64,
+        61,
+        61,
+        61,
+        1382,
+        61,
+        50,
+        406,
+        61,
+        50,
+        61,
+        61,
+        60,
+        2319,
+        61,
+        61
+      ],
+      "decoded_action_hist": {},
+      "fixed_episode_seeds": true,
+      "eval_seed": 42,
+      "episode_seeds": [
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null
+      ],
+      "episode_indices": [
+        0,
+        1,
+        2,
+        3,
+        4,
+        5,
+        6,
+        7,
+        8,
+        9,
+        10,
+        11,
+        12,
+        13,
+        14,
+        15,
+        16,
+        17,
+        18,
+        19
+      ]
+    }
+  },
+  "aggregate": {
+    "stage": "mid_train",
+    "step": 1000,
+    "task": "flappy",
+    "model_alias": "openvla",
+    "fixed_episode_seeds": true,
+    "eval_seed": 42,
+    "total_episodes": 20,
+    "mean_reward": 29.395000371709465,
+    "mean_length": 255.9,
+    "std_reward": 69.04561960432308,
+    "std_length": 556.9244024102372,
+    "task_count": 1,
+    "macro_mean_reward": 29.395000371709465,
+    "macro_mean_length": 255.9,
+    "distributed_eval": true
+  }
+}

flappy_fix_latency_2_200ep_7k2steps_plain_multi/eval/mid_train/step_1250.json

@@ -0,0 +1,120 @@
+{
+  "per_latency": {
+    "flappy/latency_2": {
+      "latency": 2,
+      "num_episodes": 20,
+      "mean_reward": 253.63500299341976,
+      "mean_length": 2063.0,
+      "std_reward": 160.78951731733707,
+      "std_length": 1294.759552967268,
+      "episode_rewards": [
+        118.70000141113997,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        102.90000123530626,
+        244.6000028848648,
+        444.6000052243471,
+        47.00000058114529,
+        426.60000501573086,
+        149.20000177621841,
+        444.6000052243471,
+        107.30000128597021,
+        218.40000258386135,
+        76.50000091642141,
+        444.6000052243471,
+        32.80000041425228,
+        23.200000301003456,
+        228.700002707541,
+        384.5000045225024,
+        244.70000288635492
+      ],
+      "episode_lengths": [
+        973,
+        3600,
+        3600,
+        3600,
+        851,
+        1989,
+        3600,
+        400,
+        3458,
+        1224,
+        3600,
+        886,
+        1781,
+        632,
+        3600,
+        285,
+        207,
+        1866,
+        3118,
+        1990
+      ],
+      "decoded_action_hist": {},
+      "fixed_episode_seeds": true,
+      "eval_seed": 42,
+      "episode_seeds": [
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null
+      ],
+      "episode_indices": [
+        0,
+        1,
+        2,
+        3,
+        4,
+        5,
+        6,
+        7,
+        8,
+        9,
+        10,
+        11,
+        12,
+        13,
+        14,
+        15,
+        16,
+        17,
+        18,
+        19
+      ]
+    }
+  },
+  "aggregate": {
+    "stage": "mid_train",
+    "step": 1250,
+    "task": "flappy",
+    "model_alias": "openvla",
+    "fixed_episode_seeds": true,
+    "eval_seed": 42,
+    "total_episodes": 20,
+    "mean_reward": 253.63500299341976,
+    "mean_length": 2063.0,
+    "std_reward": 160.78951731733707,
+    "std_length": 1294.759552967268,
+    "task_count": 1,
+    "macro_mean_reward": 253.63500299341976,
+    "macro_mean_length": 2063.0,
+    "distributed_eval": true
+  }
+}

flappy_fix_latency_2_200ep_7k2steps_plain_multi/eval/mid_train/step_1500.json

@@ -0,0 +1,120 @@
+{
+  "per_latency": {
+    "flappy/latency_2": {
+      "latency": 2,
+      "num_episodes": 20,
+      "mean_reward": 349.4700041130185,
+      "mean_length": 2834.35,
+      "std_reward": 147.5672443637608,
+      "std_length": 1188.609829801184,
+      "episode_rewards": [
+        134.4000016003847,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        325.9000038430095,
+        444.6000052243471,
+        208.90000247210264,
+        444.6000052243471,
+        444.6000052243471,
+        13.000000178813934,
+        353.900004170835,
+        90.6000010818243,
+        444.6000052243471,
+        444.6000052243471,
+        82.90000099688768,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471
+      ],
+      "episode_lengths": [
+        1103,
+        3600,
+        3600,
+        3600,
+        2649,
+        3600,
+        1704,
+        3600,
+        3600,
+        123,
+        2875,
+        746,
+        3600,
+        3600,
+        687,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600
+      ],
+      "decoded_action_hist": {},
+      "fixed_episode_seeds": true,
+      "eval_seed": 42,
+      "episode_seeds": [
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null
+      ],
+      "episode_indices": [
+        0,
+        1,
+        2,
+        3,
+        4,
+        5,
+        6,
+        7,
+        8,
+        9,
+        10,
+        11,
+        12,
+        13,
+        14,
+        15,
+        16,
+        17,
+        18,
+        19
+      ]
+    }
+  },
+  "aggregate": {
+    "stage": "mid_train",
+    "step": 1500,
+    "task": "flappy",
+    "model_alias": "openvla",
+    "fixed_episode_seeds": true,
+    "eval_seed": 42,
+    "total_episodes": 20,
+    "mean_reward": 349.4700041130185,
+    "mean_length": 2834.35,
+    "std_reward": 147.5672443637608,
+    "std_length": 1188.609829801184,
+    "task_count": 1,
+    "macro_mean_reward": 349.4700041130185,
+    "macro_mean_length": 2834.35,
+    "distributed_eval": true
+  }
+}

flappy_fix_latency_2_200ep_7k2steps_plain_multi/eval/mid_train/step_1750.json

@@ -0,0 +1,120 @@
+{
+  "per_latency": {
+    "flappy/latency_2": {
+      "latency": 2,
+      "num_episodes": 20,
+      "mean_reward": 256.01500301845374,
+      "mean_length": 2080.5,
+      "std_reward": 161.3821578877714,
+      "std_length": 1301.248304513785,
+      "episode_rewards": [
+        379.5000044628978,
+        113.50000134855509,
+        132.60000157356262,
+        76.80000092089176,
+        444.6000052243471,
+        444.6000052243471,
+        62.80000075697899,
+        47.00000058114529,
+        151.00000178813934,
+        444.6000052243471,
+        444.6000052243471,
+        109.00000129640102,
+        90.6000010818243,
+        387.3000045642257,
+        361.00000424683094,
+        309.5000036433339,
+        444.6000052243471,
+        444.6000052243471,
+        34.60000042617321,
+        197.50000233203173
+      ],
+      "episode_lengths": [
+        3077,
+        930,
+        1085,
+        635,
+        3600,
+        3600,
+        522,
+        400,
+        1233,
+        3600,
+        3600,
+        894,
+        746,
+        3146,
+        2928,
+        2512,
+        3600,
+        3600,
+        294,
+        1608
+      ],
+      "decoded_action_hist": {},
+      "fixed_episode_seeds": true,
+      "eval_seed": 42,
+      "episode_seeds": [
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null
+      ],
+      "episode_indices": [
+        0,
+        1,
+        2,
+        3,
+        4,
+        5,
+        6,
+        7,
+        8,
+        9,
+        10,
+        11,
+        12,
+        13,
+        14,
+        15,
+        16,
+        17,
+        18,
+        19
+      ]
+    }
+  },
+  "aggregate": {
+    "stage": "mid_train",
+    "step": 1750,
+    "task": "flappy",
+    "model_alias": "openvla",
+    "fixed_episode_seeds": true,
+    "eval_seed": 42,
+    "total_episodes": 20,
+    "mean_reward": 256.01500301845374,
+    "mean_length": 2080.5,
+    "std_reward": 161.3821578877714,
+    "std_length": 1301.248304513785,
+    "task_count": 1,
+    "macro_mean_reward": 256.01500301845374,
+    "macro_mean_length": 2080.5,
+    "distributed_eval": true
+  }
+}

flappy_fix_latency_2_200ep_7k2steps_plain_multi/eval/mid_train/step_2000.json

@@ -0,0 +1,120 @@
+{
+  "per_latency": {
+    "flappy/latency_2": {
+      "latency": 2,
+      "num_episodes": 20,
+      "mean_reward": 411.61500483863057,
+      "mean_length": 3334.25,
+      "std_reward": 101.14185342523868,
+      "std_length": 815.0359424589814,
+      "episode_rewards": [
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        48.60000059008598,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        180.9000021442771
+      ],
+      "episode_lengths": [
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        407,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        1478
+      ],
+      "decoded_action_hist": {},
+      "fixed_episode_seeds": true,
+      "eval_seed": 42,
+      "episode_seeds": [
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null
+      ],
+      "episode_indices": [
+        0,
+        1,
+        2,
+        3,
+        4,
+        5,
+        6,
+        7,
+        8,
+        9,
+        10,
+        11,
+        12,
+        13,
+        14,
+        15,
+        16,
+        17,
+        18,
+        19
+      ]
+    }
+  },
+  "aggregate": {
+    "stage": "mid_train",
+    "step": 2000,
+    "task": "flappy",
+    "model_alias": "openvla",
+    "fixed_episode_seeds": true,
+    "eval_seed": 42,
+    "total_episodes": 20,
+    "mean_reward": 411.61500483863057,
+    "mean_length": 3334.25,
+    "std_reward": 101.14185342523868,
+    "std_length": 815.0359424589814,
+    "task_count": 1,
+    "macro_mean_reward": 411.61500483863057,
+    "macro_mean_length": 3334.25,
+    "distributed_eval": true
+  }
+}

flappy_fix_latency_2_200ep_7k2steps_plain_multi/eval/mid_train/step_2250.json

@@ -0,0 +1,120 @@
+{
+  "per_latency": {
+    "flappy/latency_2": {
+      "latency": 2,
+      "num_episodes": 20,
+      "mean_reward": 444.6000052243471,
+      "mean_length": 3600.0,
+      "std_reward": 0.0,
+      "std_length": 0.0,
+      "episode_rewards": [
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471
+      ],
+      "episode_lengths": [
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600
+      ],
+      "decoded_action_hist": {},
+      "fixed_episode_seeds": true,
+      "eval_seed": 42,
+      "episode_seeds": [
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null
+      ],
+      "episode_indices": [
+        0,
+        1,
+        2,
+        3,
+        4,
+        5,
+        6,
+        7,
+        8,
+        9,
+        10,
+        11,
+        12,
+        13,
+        14,
+        15,
+        16,
+        17,
+        18,
+        19
+      ]
+    }
+  },
+  "aggregate": {
+    "stage": "mid_train",
+    "step": 2250,
+    "task": "flappy",
+    "model_alias": "openvla",
+    "fixed_episode_seeds": true,
+    "eval_seed": 42,
+    "total_episodes": 20,
+    "mean_reward": 444.6000052243471,
+    "mean_length": 3600.0,
+    "std_reward": 0.0,
+    "std_length": 0.0,
+    "task_count": 1,
+    "macro_mean_reward": 444.6000052243471,
+    "macro_mean_length": 3600.0,
+    "distributed_eval": true
+  }
+}

flappy_fix_latency_2_200ep_7k2steps_plain_multi/eval/mid_train/step_250.json

@@ -0,0 +1,120 @@
+{
+  "per_latency": {
+    "flappy/latency_2": {
+      "latency": 2,
+      "num_episodes": 20,
+      "mean_reward": 6.000000090152025,
+      "mean_length": 62.45,
+      "std_reward": 0.2792848018357637,
+      "std_length": 0.9205976319760986,
+      "episode_rewards": [
+        6.100000090897083,
+        6.000000089406967,
+        4.800000086426735,
+        6.100000090897083,
+        6.000000089406967,
+        6.100000090897083,
+        6.100000090897083,
+        6.100000090897083,
+        6.000000089406967,
+        6.100000090897083,
+        6.100000090897083,
+        6.000000089406967,
+        6.100000090897083,
+        6.100000090897083,
+        6.100000090897083,
+        6.100000090897083,
+        6.000000089406967,
+        6.000000089406967,
+        6.000000089406967,
+        6.100000090897083
+      ],
+      "episode_lengths": [
+        63,
+        62,
+        59,
+        63,
+        62,
+        63,
+        63,
+        63,
+        62,
+        63,
+        63,
+        62,
+        63,
+        63,
+        63,
+        63,
+        62,
+        62,
+        62,
+        63
+      ],
+      "decoded_action_hist": {},
+      "fixed_episode_seeds": true,
+      "eval_seed": 42,
+      "episode_seeds": [
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null
+      ],
+      "episode_indices": [
+        0,
+        1,
+        2,
+        3,
+        4,
+        5,
+        6,
+        7,
+        8,
+        9,
+        10,
+        11,
+        12,
+        13,
+        14,
+        15,
+        16,
+        17,
+        18,
+        19
+      ]
+    }
+  },
+  "aggregate": {
+    "stage": "mid_train",
+    "step": 250,
+    "task": "flappy",
+    "model_alias": "openvla",
+    "fixed_episode_seeds": true,
+    "eval_seed": 42,
+    "total_episodes": 20,
+    "mean_reward": 6.000000090152025,
+    "mean_length": 62.45,
+    "std_reward": 0.2792848018357637,
+    "std_length": 0.9205976319760986,
+    "task_count": 1,
+    "macro_mean_reward": 6.000000090152025,
+    "macro_mean_length": 62.45,
+    "distributed_eval": true
+  }
+}

flappy_fix_latency_2_200ep_7k2steps_plain_multi/eval/mid_train/step_2500.json

@@ -0,0 +1,120 @@
+{
+  "per_latency": {
+    "flappy/latency_2": {
+      "latency": 2,
+      "num_episodes": 20,
+      "mean_reward": 354.9100041769445,
+      "mean_length": 2878.4,
+      "std_reward": 139.38672588911118,
+      "std_length": 1122.3577148128843,
+      "episode_rewards": [
+        228.60000270605087,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        46.60000057518482,
+        353.900004170835,
+        444.6000052243471,
+        444.6000052243471,
+        344.4000040590763,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        121.30000144988298,
+        444.6000052243471,
+        104.80000124871731,
+        118.80000141263008
+      ],
+      "episode_lengths": [
+        1865,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        396,
+        2875,
+        3600,
+        3600,
+        2798,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        999,
+        3600,
+        861,
+        974
+      ],
+      "decoded_action_hist": {},
+      "fixed_episode_seeds": true,
+      "eval_seed": 42,
+      "episode_seeds": [
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null
+      ],
+      "episode_indices": [
+        0,
+        1,
+        2,
+        3,
+        4,
+        5,
+        6,
+        7,
+        8,
+        9,
+        10,
+        11,
+        12,
+        13,
+        14,
+        15,
+        16,
+        17,
+        18,
+        19
+      ]
+    }
+  },
+  "aggregate": {
+    "stage": "mid_train",
+    "step": 2500,
+    "task": "flappy",
+    "model_alias": "openvla",
+    "fixed_episode_seeds": true,
+    "eval_seed": 42,
+    "total_episodes": 20,
+    "mean_reward": 354.9100041769445,
+    "mean_length": 2878.4,
+    "std_reward": 139.38672588911118,
+    "std_length": 1122.3577148128843,
+    "task_count": 1,
+    "macro_mean_reward": 354.9100041769445,
+    "macro_mean_length": 2878.4,
+    "distributed_eval": true
+  }
+}

flappy_fix_latency_2_200ep_7k2steps_plain_multi/eval/mid_train/step_2750.json

@@ -0,0 +1,120 @@
+{
+  "per_latency": {
+    "flappy/latency_2": {
+      "latency": 2,
+      "num_episodes": 20,
+      "mean_reward": 426.8100050166249,
+      "mean_length": 3457.05,
+      "std_reward": 70.95351307600934,
+      "std_length": 572.0222438856725,
+      "episode_rewards": [
+        444.6000052243471,
+        118.50000140815973,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        440.60000517964363,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        418.90000493079424,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471
+      ],
+      "episode_lengths": [
+        3600,
+        971,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3571,
+        3600,
+        3600,
+        3600,
+        3399,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600
+      ],
+      "decoded_action_hist": {},
+      "fixed_episode_seeds": true,
+      "eval_seed": 42,
+      "episode_seeds": [
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null
+      ],
+      "episode_indices": [
+        0,
+        1,
+        2,
+        3,
+        4,
+        5,
+        6,
+        7,
+        8,
+        9,
+        10,
+        11,
+        12,
+        13,
+        14,
+        15,
+        16,
+        17,
+        18,
+        19
+      ]
+    }
+  },
+  "aggregate": {
+    "stage": "mid_train",
+    "step": 2750,
+    "task": "flappy",
+    "model_alias": "openvla",
+    "fixed_episode_seeds": true,
+    "eval_seed": 42,
+    "total_episodes": 20,
+    "mean_reward": 426.8100050166249,
+    "mean_length": 3457.05,
+    "std_reward": 70.95351307600934,
+    "std_length": 572.0222438856725,
+    "task_count": 1,
+    "macro_mean_reward": 426.8100050166249,
+    "macro_mean_length": 3457.05,
+    "distributed_eval": true
+  }
+}

flappy_fix_latency_2_200ep_7k2steps_plain_multi/eval/mid_train/step_3000.json

@@ -0,0 +1,120 @@
+{
+  "per_latency": {
+    "flappy/latency_2": {
+      "latency": 2,
+      "num_episodes": 20,
+      "mean_reward": 422.8200049698353,
+      "mean_length": 3424.7,
+      "std_reward": 73.7419604241364,
+      "std_length": 594.1030297852384,
+      "episode_rewards": [
+        444.6000052243471,
+        118.70000141113997,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        334.9000039473176,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471
+      ],
+      "episode_lengths": [
+        3600,
+        973,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        2721,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600
+      ],
+      "decoded_action_hist": {},
+      "fixed_episode_seeds": true,
+      "eval_seed": 42,
+      "episode_seeds": [
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null
+      ],
+      "episode_indices": [
+        0,
+        1,
+        2,
+        3,
+        4,
+        5,
+        6,
+        7,
+        8,
+        9,
+        10,
+        11,
+        12,
+        13,
+        14,
+        15,
+        16,
+        17,
+        18,
+        19
+      ]
+    }
+  },
+  "aggregate": {
+    "stage": "mid_train",
+    "step": 3000,
+    "task": "flappy",
+    "model_alias": "openvla",
+    "fixed_episode_seeds": true,
+    "eval_seed": 42,
+    "total_episodes": 20,
+    "mean_reward": 422.8200049698353,
+    "mean_length": 3424.7,
+    "std_reward": 73.7419604241364,
+    "std_length": 594.1030297852384,
+    "task_count": 1,
+    "macro_mean_reward": 422.8200049698353,
+    "macro_mean_length": 3424.7,
+    "distributed_eval": true
+  }
+}

flappy_fix_latency_2_200ep_7k2steps_plain_multi/eval/mid_train/step_3250.json

@@ -0,0 +1,120 @@
+{
+  "per_latency": {
+    "flappy/latency_2": {
+      "latency": 2,
+      "num_episodes": 20,
+      "mean_reward": 417.7050049100071,
+      "mean_length": 3383.45,
+      "std_reward": 96.75278649231183,
+      "std_length": 779.8917537068845,
+      "episode_rewards": [
+        444.6000052243471,
+        444.6000052243471,
+        6.800000101327896,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        344.5000040605664,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471
+      ],
+      "episode_lengths": [
+        3600,
+        3600,
+        70,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        2799,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600
+      ],
+      "decoded_action_hist": {},
+      "fixed_episode_seeds": true,
+      "eval_seed": 42,
+      "episode_seeds": [
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null
+      ],
+      "episode_indices": [
+        0,
+        1,
+        2,
+        3,
+        4,
+        5,
+        6,
+        7,
+        8,
+        9,
+        10,
+        11,
+        12,
+        13,
+        14,
+        15,
+        16,
+        17,
+        18,
+        19
+      ]
+    }
+  },
+  "aggregate": {
+    "stage": "mid_train",
+    "step": 3250,
+    "task": "flappy",
+    "model_alias": "openvla",
+    "fixed_episode_seeds": true,
+    "eval_seed": 42,
+    "total_episodes": 20,
+    "mean_reward": 417.7050049100071,
+    "mean_length": 3383.45,
+    "std_reward": 96.75278649231183,
+    "std_length": 779.8917537068845,
+    "task_count": 1,
+    "macro_mean_reward": 417.7050049100071,
+    "macro_mean_length": 3383.45,
+    "distributed_eval": true
+  }
+}

flappy_fix_latency_2_200ep_7k2steps_plain_multi/eval/mid_train/step_3500.json

@@ -0,0 +1,120 @@
+{
+  "per_latency": {
+    "flappy/latency_2": {
+      "latency": 2,
+      "num_episodes": 20,
+      "mean_reward": 412.7750048529357,
+      "mean_length": 3344.25,
+      "std_reward": 80.25203440101174,
+      "std_length": 645.9581159022619,
+      "episode_rewards": [
+        444.6000052243471,
+        120.50000143796206,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        334.9000039473176,
+        267.6000031530857,
+        418.90000493079424,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471
+      ],
+      "episode_lengths": [
+        3600,
+        991,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        2721,
+        2174,
+        3399,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600
+      ],
+      "decoded_action_hist": {},
+      "fixed_episode_seeds": true,
+      "eval_seed": 42,
+      "episode_seeds": [
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null
+      ],
+      "episode_indices": [
+        0,
+        1,
+        2,
+        3,
+        4,
+        5,
+        6,
+        7,
+        8,
+        9,
+        10,
+        11,
+        12,
+        13,
+        14,
+        15,
+        16,
+        17,
+        18,
+        19
+      ]
+    }
+  },
+  "aggregate": {
+    "stage": "mid_train",
+    "step": 3500,
+    "task": "flappy",
+    "model_alias": "openvla",
+    "fixed_episode_seeds": true,
+    "eval_seed": 42,
+    "total_episodes": 20,
+    "mean_reward": 412.7750048529357,
+    "mean_length": 3344.25,
+    "std_reward": 80.25203440101174,
+    "std_length": 645.9581159022619,
+    "task_count": 1,
+    "macro_mean_reward": 412.7750048529357,
+    "macro_mean_length": 3344.25,
+    "distributed_eval": true
+  }
+}

flappy_fix_latency_2_200ep_7k2steps_plain_multi/eval/mid_train/step_3750.json

@@ -0,0 +1,120 @@
+{
+  "per_latency": {
+    "flappy/latency_2": {
+      "latency": 2,
+      "num_episodes": 20,
+      "mean_reward": 414.9000048771501,
+      "mean_length": 3360.8,
+      "std_reward": 89.51117911923446,
+      "std_length": 720.8735395338076,
+      "episode_rewards": [
+        444.6000052243471,
+        120.50000143796206,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        174.70000206679106,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471
+      ],
+      "episode_lengths": [
+        3600,
+        991,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        1425,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600
+      ],
+      "decoded_action_hist": {},
+      "fixed_episode_seeds": true,
+      "eval_seed": 42,
+      "episode_seeds": [
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null
+      ],
+      "episode_indices": [
+        0,
+        1,
+        2,
+        3,
+        4,
+        5,
+        6,
+        7,
+        8,
+        9,
+        10,
+        11,
+        12,
+        13,
+        14,
+        15,
+        16,
+        17,
+        18,
+        19
+      ]
+    }
+  },
+  "aggregate": {
+    "stage": "mid_train",
+    "step": 3750,
+    "task": "flappy",
+    "model_alias": "openvla",
+    "fixed_episode_seeds": true,
+    "eval_seed": 42,
+    "total_episodes": 20,
+    "mean_reward": 414.9000048771501,
+    "mean_length": 3360.8,
+    "std_reward": 89.51117911923446,
+    "std_length": 720.8735395338076,
+    "task_count": 1,
+    "macro_mean_reward": 414.9000048771501,
+    "macro_mean_length": 3360.8,
+    "distributed_eval": true
+  }
+}

flappy_fix_latency_2_200ep_7k2steps_plain_multi/eval/mid_train/step_4000.json

@@ -0,0 +1,120 @@
+{
+  "per_latency": {
+    "flappy/latency_2": {
+      "latency": 2,
+      "num_episodes": 20,
+      "mean_reward": 408.3500048011541,
+      "mean_length": 3308.45,
+      "std_reward": 93.69927803918887,
+      "std_length": 753.9609721331734,
+      "episode_rewards": [
+        444.6000052243471,
+        120.50000143796206,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        334.9000039473176,
+        444.6000052243471,
+        444.6000052243471,
+        153.40000182390213,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471
+      ],
+      "episode_lengths": [
+        3600,
+        991,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        2721,
+        3600,
+        3600,
+        1257,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600
+      ],
+      "decoded_action_hist": {},
+      "fixed_episode_seeds": true,
+      "eval_seed": 42,
+      "episode_seeds": [
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null
+      ],
+      "episode_indices": [
+        0,
+        1,
+        2,
+        3,
+        4,
+        5,
+        6,
+        7,
+        8,
+        9,
+        10,
+        11,
+        12,
+        13,
+        14,
+        15,
+        16,
+        17,
+        18,
+        19
+      ]
+    }
+  },
+  "aggregate": {
+    "stage": "mid_train",
+    "step": 4000,
+    "task": "flappy",
+    "model_alias": "openvla",
+    "fixed_episode_seeds": true,
+    "eval_seed": 42,
+    "total_episodes": 20,
+    "mean_reward": 408.3500048011541,
+    "mean_length": 3308.45,
+    "std_reward": 93.69927803918887,
+    "std_length": 753.9609721331734,
+    "task_count": 1,
+    "macro_mean_reward": 408.3500048011541,
+    "macro_mean_length": 3308.45,
+    "distributed_eval": true
+  }
+}

flappy_fix_latency_2_200ep_7k2steps_plain_multi/eval/mid_train/step_500.json

@@ -0,0 +1,120 @@
+{
+  "per_latency": {
+    "flappy/latency_2": {
+      "latency": 2,
+      "num_episodes": 20,
+      "mean_reward": 13.590000185370446,
+      "mean_length": 127.55,
+      "std_reward": 31.173288608948123,
+      "std_length": 251.05725940510064,
+      "episode_rewards": [
+        4.800000086426735,
+        32.700000412762165,
+        3.90000007301569,
+        3.90000007301569,
+        146.80000174045563,
+        4.800000086426735,
+        3.90000007301569,
+        4.800000086426735,
+        6.3000000938773155,
+        6.600000098347664,
+        3.90000007301569,
+        6.3000000938773155,
+        3.90000007301569,
+        3.90000007301569,
+        4.500000081956387,
+        4.800000086426735,
+        9.200000137090683,
+        6.3000000938773155,
+        3.90000007301569,
+        6.600000098347664
+      ],
+      "episode_lengths": [
+        59,
+        284,
+        50,
+        50,
+        1200,
+        59,
+        50,
+        59,
+        65,
+        68,
+        50,
+        65,
+        50,
+        50,
+        56,
+        59,
+        94,
+        65,
+        50,
+        68
+      ],
+      "decoded_action_hist": {},
+      "fixed_episode_seeds": true,
+      "eval_seed": 42,
+      "episode_seeds": [
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null
+      ],
+      "episode_indices": [
+        0,
+        1,
+        2,
+        3,
+        4,
+        5,
+        6,
+        7,
+        8,
+        9,
+        10,
+        11,
+        12,
+        13,
+        14,
+        15,
+        16,
+        17,
+        18,
+        19
+      ]
+    }
+  },
+  "aggregate": {
+    "stage": "mid_train",
+    "step": 500,
+    "task": "flappy",
+    "model_alias": "openvla",
+    "fixed_episode_seeds": true,
+    "eval_seed": 42,
+    "total_episodes": 20,
+    "mean_reward": 13.590000185370446,
+    "mean_length": 127.55,
+    "std_reward": 31.173288608948123,
+    "std_length": 251.05725940510064,
+    "task_count": 1,
+    "macro_mean_reward": 13.590000185370446,
+    "macro_mean_length": 127.55,
+    "distributed_eval": true
+  }
+}

flappy_fix_latency_2_200ep_7k2steps_plain_multi/eval/mid_train/step_750.json

@@ -0,0 +1,120 @@
+{
+  "per_latency": {
+    "flappy/latency_2": {
+      "latency": 2,
+      "num_episodes": 20,
+      "mean_reward": 38.45000047534704,
+      "mean_length": 327.45,
+      "std_reward": 105.02031592997427,
+      "std_length": 846.6170016601368,
+      "episode_rewards": [
+        6.500000096857548,
+        444.6000052243471,
+        4.500000081956387,
+        6.200000092387199,
+        8.400000125169754,
+        6.500000096857548,
+        3.90000007301569,
+        6.500000096857548,
+        5.000000089406967,
+        4.100000075995922,
+        6.200000092387199,
+        4.900000087916851,
+        4.100000075995922,
+        6.200000092387199,
+        4.100000075995922,
+        6.500000096857548,
+        227.90000269562006,
+        4.900000087916851,
+        3.90000007301569,
+        4.100000075995922
+      ],
+      "episode_lengths": [
+        67,
+        3600,
+        56,
+        64,
+        86,
+        67,
+        50,
+        67,
+        61,
+        52,
+        64,
+        60,
+        52,
+        64,
+        52,
+        67,
+        1858,
+        60,
+        50,
+        52
+      ],
+      "decoded_action_hist": {},
+      "fixed_episode_seeds": true,
+      "eval_seed": 42,
+      "episode_seeds": [
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null
+      ],
+      "episode_indices": [
+        0,
+        1,
+        2,
+        3,
+        4,
+        5,
+        6,
+        7,
+        8,
+        9,
+        10,
+        11,
+        12,
+        13,
+        14,
+        15,
+        16,
+        17,
+        18,
+        19
+      ]
+    }
+  },
+  "aggregate": {
+    "stage": "mid_train",
+    "step": 750,
+    "task": "flappy",
+    "model_alias": "openvla",
+    "fixed_episode_seeds": true,
+    "eval_seed": 42,
+    "total_episodes": 20,
+    "mean_reward": 38.45000047534704,
+    "mean_length": 327.45,
+    "std_reward": 105.02031592997427,
+    "std_length": 846.6170016601368,
+    "task_count": 1,
+    "macro_mean_reward": 38.45000047534704,
+    "macro_mean_length": 327.45,
+    "distributed_eval": true
+  }
+}

flappy_fix_latency_2_200ep_7k2steps_plain_multi/eval/post_train/step_4000.json

@@ -0,0 +1,240 @@
+{
+  "per_latency": {
+    "flappy/latency_2": {
+      "latency": 2,
+      "num_episodes": 50,
+      "mean_reward": 403.98200474992393,
+      "mean_length": 3273.12,
+      "std_reward": 109.87717942362134,
+      "std_length": 884.5486225188529,
+      "episode_rewards": [
+        444.6000052243471,
+        120.50000143796206,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        334.9000039473176,
+        444.6000052243471,
+        444.6000052243471,
+        153.40000182390213,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        4.700000084936619,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        4.700000084936619,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        174.70000206679106,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        288.4000034034252,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471
+      ],
+      "episode_lengths": [
+        3600,
+        991,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        2721,
+        3600,
+        3600,
+        1257,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        58,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        58,
+        3600,
+        3600,
+        3600,
+        3600,
+        1425,
+        3600,
+        3600,
+        3600,
+        2346,
+        3600,
+        3600,
+        3600,
+        3600
+      ],
+      "decoded_action_hist": {},
+      "fixed_episode_seeds": true,
+      "eval_seed": 42,
+      "episode_seeds": [
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null
+      ],
+      "episode_indices": [
+        0,
+        1,
+        2,
+        3,
+        4,
+        5,
+        6,
+        7,
+        8,
+        9,
+        10,
+        11,
+        12,
+        13,
+        14,
+        15,
+        16,
+        17,
+        18,
+        19,
+        20,
+        21,
+        22,
+        23,
+        24,
+        25,
+        26,
+        27,
+        28,
+        29,
+        30,
+        31,
+        32,
+        33,
+        34,
+        35,
+        36,
+        37,
+        38,
+        39,
+        40,
+        41,
+        42,
+        43,
+        44,
+        45,
+        46,
+        47,
+        48,
+        49
+      ]
+    }
+  },
+  "aggregate": {
+    "stage": "post_train",
+    "step": 4000,
+    "task": "flappy",
+    "model_alias": "openvla",
+    "fixed_episode_seeds": true,
+    "eval_seed": 42,
+    "total_episodes": 50,
+    "mean_reward": 403.98200474992393,
+    "mean_length": 3273.12,
+    "std_reward": 109.87717942362134,
+    "std_length": 884.5486225188529,
+    "task_count": 1,
+    "macro_mean_reward": 403.98200474992393,
+    "macro_mean_length": 3273.12,
+    "distributed_eval": true
+  }
+}

flappy_fix_latency_2_200ep_7k2steps_plain_multi/hydra/.hydra/config.yaml

@@ -0,0 +1,247 @@
+framework:
+  qwenvl:
+    base_vlm: /workspace/latency-sensitive-bench/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct
+    attn_implementation: flash_attention_2
+    flex_backend: triton
+    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
+  kv_memory:
+    enabled: false
+    window: 4
+    rollout_len: 8
+    packed_train: false
+    rebased_sink: true
+  name: QwenOFT
+datasets:
+  vla_data:
+    dataset_py: lerobot_datasets
+    include_state: true
+    data_root_dir: /workspace/latency-sensitive-bench/starVLA/data/flappy_fix_latency_2_200ep_7k2steps
+    data_mix: flappy_train__bridge
+    eval_data_mix: flappy_train__bridge__val
+    custom_mixtures_path: null
+    action_type: discrete
+    sequential_step_sampling: false
+    eval_sequential_step_sampling: null
+    num_workers: 8
+    eval_num_workers: 8
+    prefetch_factor: 4
+    persistent_workers: true
+    pin_memory: true
+    shuffle: true
+    action_balance:
+      enabled: false
+      strategy: balanced_epoch
+      action_key: action_id
+      target_flap_fraction: 0.3
+      noop_id: 0
+      flap_id: 1
+    latency_curriculum:
+      enabled: false
+      strategy: exclusive
+      latencies: null
+      phase_steps: null
+      phase_distributions: null
+      new_latency_passes: 1.0
+      replay_passes: 0.25
+      target_total_passes: 2.0
+      final_equalization: true
+      step_budget_mode: auto
+      eval_at_phase_end: false
+      save_at_phase_end: false
+      computed_plan: null
+    per_device_batch_size: 16
+    load_all_data_for_training: true
+    num_obs_frames: 4
+    image_mode: multiframe
+    prompt_mode: raw
+    stitch_grid:
+    - 2
+    - 2
+    obs_image_size: null
+    video_backend: torchvision_av
+dataset:
+  source_hf: ''
+  config_name: null
+  source_subdir: null
+  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
+  skip_verification: 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: 4000
+  num_warmup_steps: 100
+  save_interval: 500
+  eval_interval: 250
+  eval_num_batches: 50
+  per_latency_eval_num_batches: null
+  eval_action_classification: false
+  eval_action_classification_interval: null
+  cc_f1_tolerance: 1
+  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: ''
+  freeze_vit: false
+  freeze_tied_embedding: false
+  freeze_llm_layers: []
+  loss_scale:
+    vla: 1.0
+    vlm: 0.1
+  max_grad_norm: 1.0
+  weight_decay: 0.0
+  logging_frequency: 1
+  profile_timing:
+    enabled: false
+    log_interval: 10
+  gradient_clipping: 1.0
+  gradient_accumulation_steps: 4
+  distributed_backend: deepspeed
+  is_resume: false
+  pretrained_checkpoint: /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/checkpoints/_initialization/StarVLA_Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt
+  resume_step: 0
+  reload_modules: null
+  optimizer:
+    name: AdamW
+    betas:
+    - 0.9
+    - 0.95
+    eps: 1.0e-08
+    weight_decay: 1.0e-08
+    fused: true
+  save_format: pt
+workspace_dir: WORKSPACE_DIR
+run_root_dir: /workspace/latency-sensitive-bench/starVLA/results/Checkpoints
+seed: 42
+wandb_entity: saberrr-zju
+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: data/flappy_fix_latency_2_200ep_7k2steps
+  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
+    image_transform: raw_rgb
+    prompt_mode: raw
+    ghost_trail:
+      history_frames: 5
+      gamma: 1.3
+      min_alpha: 35
+      scroll_px_per_step: 4.0
+      ground_fraction: 0.22
+    seed: 42
+    fixed_episode_seeds: true
+    latency_seed_stride: 0
+    task_seed_stride: 0
+    task_description: ''
+    eval_parallel_envs: 5
+    enabled: true
+    eval_backend: latency_bench
+    distributed_mode: rank_sharded
+    vectorized:
+      enabled: false
+      batch_size: 1
+    latency:
+      prompt_map_path: /workspace/latency-sensitive-bench/starVLA/data/flappy_fix_latency_2_200ep_7k2steps/flappy_train__bridge/latency_prompt_map.json
+      mode: single
+      values:
+      - 0
+    mid_train:
+      enabled: true
+      interval_steps: 250
+      latencies:
+      - 2
+      num_episodes: 20
+      max_steps_per_episode: 3600
+    post_train:
+      enabled: true
+      latencies:
+      - 2
+      num_episodes: 50
+      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: none
+  hf_repo_id: null
+  save_best_model: false
+  save_final_model: true
+  save_pt_file: false
+  save_safetensors_file: true
+  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: flappy_fix_latency_2_200ep_7k2steps_plain_multi
+output_dir: null
+config_yaml: null
+is_debug: false
+version_id: 0.21

flappy_fix_latency_2_200ep_7k2steps_plain_multi/hydra/.hydra/hydra.yaml

@@ -0,0 +1,368 @@
+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
+    - ++framework.qwenvl.base_vlm=playground/Pretrained_models/Qwen3-VL-4B-Instruct
+    - ++framework.qwenvl.attn_implementation=flash_attention_2
+    - ++framework.qwenvl.flex_backend=triton
+    - ++framework.qwenvl.enable_gradient_checkpointing=true
+    - ++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
+    - ++framework.action_model.action_dim=7
+    - ++framework.action_model.action_env_dim=2
+    - ++framework.kv_memory.enabled=false
+    - ++framework.kv_memory.window=4
+    - ++framework.kv_memory.rollout_len=8
+    - ++framework.kv_memory.packed_train=false
+    - ++framework.kv_memory.rebased_sink=true
+    - ++framework.name=QwenOFT
+    - ++datasets.vla_data.dataset_py=lerobot_datasets
+    - ++datasets.vla_data.include_state=true
+    - ++datasets.vla_data.data_root_dir=playground/Datasets/rl_games
+    - ++datasets.vla_data.data_mix=flappy_train
+    - ++datasets.vla_data.eval_data_mix=null
+    - ++datasets.vla_data.custom_mixtures_path=null
+    - ++datasets.vla_data.action_type=discrete
+    - ++datasets.vla_data.sequential_step_sampling=false
+    - ++datasets.vla_data.eval_sequential_step_sampling=null
+    - ++datasets.vla_data.num_workers=8
+    - ++datasets.vla_data.eval_num_workers=8
+    - ++datasets.vla_data.prefetch_factor=4
+    - ++datasets.vla_data.persistent_workers=true
+    - ++datasets.vla_data.pin_memory=true
+    - ++datasets.vla_data.shuffle=true
+    - ++datasets.vla_data.action_balance.enabled=false
+    - ++datasets.vla_data.action_balance.strategy=balanced_epoch
+    - ++datasets.vla_data.action_balance.action_key=action_id
+    - ++datasets.vla_data.action_balance.target_flap_fraction=0.3
+    - ++datasets.vla_data.action_balance.noop_id=0
+    - ++datasets.vla_data.action_balance.flap_id=1
+    - ++datasets.vla_data.latency_curriculum.enabled=false
+    - ++datasets.vla_data.latency_curriculum.strategy=exclusive
+    - ++datasets.vla_data.latency_curriculum.latencies=null
+    - ++datasets.vla_data.latency_curriculum.phase_steps=null
+    - ++datasets.vla_data.latency_curriculum.phase_distributions=null
+    - ++datasets.vla_data.latency_curriculum.new_latency_passes=1.0
+    - ++datasets.vla_data.latency_curriculum.replay_passes=0.25
+    - ++datasets.vla_data.latency_curriculum.target_total_passes=2.0
+    - ++datasets.vla_data.latency_curriculum.final_equalization=true
+    - ++datasets.vla_data.latency_curriculum.step_budget_mode=auto
+    - ++datasets.vla_data.latency_curriculum.eval_at_phase_end=false
+    - ++datasets.vla_data.latency_curriculum.save_at_phase_end=false
+    - ++datasets.vla_data.latency_curriculum.computed_plan=null
+    - ++datasets.vla_data.per_device_batch_size=16
+    - ++datasets.vla_data.load_all_data_for_training=true
+    - ++datasets.vla_data.num_obs_frames=4
+    - ++datasets.vla_data.image_mode=multiframe
+    - ++datasets.vla_data.prompt_mode=raw
+    - ++datasets.vla_data.stitch_grid=[2,2]
+    - ++datasets.vla_data.obs_image_size=null
+    - ++datasets.vla_data.video_backend=torchvision_av
+    - ++dataset.source_hf=
+    - ++dataset.config_name=null
+    - ++dataset.source_subdir=null
+    - ++dataset.converted_name=flappy_train
+    - ++dataset.single_source_hf=
+    - ++dataset.mixed_source_hf=
+    - ++dataset.single_converted_name=flappy_train
+    - ++dataset.mixed_converted_name=flappy_mixed_latency_train
+    - ++dataset.single_latency_filter=null
+    - ++dataset.mixed_latency_filter=null
+    - ++dataset.force_download=false
+    - ++dataset.setup_force=false
+    - ++dataset.skip_verification=false
+    - ++dataset.verify_rows=200
+    - ++dataset.max_episodes=null
+    - ++dataset.episodes_per_latency=null
+    - ++dataset.latency_filter=null
+    - ++dataset.debug_subset.enabled=false
+    - ++dataset.debug_subset.max_episodes=5
+    - ++dataset.debug_subset.suffix=debug
+    - ++base_model.repo_id=Qwen/Qwen3-VL-4B-Instruct
+    - ++initialization.checkpoint_local_dir=playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1
+    - ++initialization.checkpoint_hf_repo_id=StarVLA/Qwen3VL-OFT-Bridge-RT-1
+    - ++initialization.checkpoint_filename=checkpoints/steps_5000_pytorch_model.pt
+    - trainer.max_train_steps=4000
+    - trainer.num_warmup_steps=100
+    - trainer.save_interval=500
+    - trainer.eval_interval=250
+    - trainer.eval_num_batches=50
+    - trainer.per_latency_eval_num_batches=null
+    - trainer.eval_action_classification=false
+    - trainer.eval_action_classification_interval=null
+    - trainer.cc_f1_tolerance=1
+    - 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.freeze_modules=
+    - trainer.freeze_vit=false
+    - trainer.freeze_tied_embedding=false
+    - trainer.freeze_llm_layers=[]
+    - trainer.loss_scale.vla=1.0
+    - trainer.loss_scale.vlm=0.1
+    - trainer.max_grad_norm=1.0
+    - trainer.weight_decay=0.0
+    - trainer.logging_frequency=1
+    - trainer.profile_timing.enabled=false
+    - trainer.profile_timing.log_interval=10
+    - trainer.gradient_clipping=1.0
+    - trainer.gradient_accumulation_steps=4
+    - trainer.distributed_backend=deepspeed
+    - trainer.is_resume=false
+    - trainer.pretrained_checkpoint=null
+    - trainer.resume_step=0
+    - trainer.reload_modules=null
+    - 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=true
+    - trainer.save_format=pt
+    - ++workspace_dir=WORKSPACE_DIR
+    - ++run_root_dir=results/Checkpoints
+    - ++seed=42
+    - ++wandb_entity=saberrr-zju
+    - ++wandb_project=starVLA_rl_games
+    - ++auth.env_file=null
+    - ++auth.hf_token_env=HF_TOKEN
+    - ++auth.wandb_api_key_env=WANDB_API_KEY
+    - ++paths.run_root_dir=results/Checkpoints
+    - ++paths.dataset_local_dir=data/flappy_fix_latency_2_200ep_7k2steps
+    - ++paths.dataset_cache_dir=null
+    - ++paths.base_model_dir=playground/Pretrained_models/Qwen3-VL-4B-Instruct
+    - ++paths.accelerate_config=starVLA/config/deepseeds/deepspeed_zero2.yaml
+    - ++rl_games.model_alias=openvla
+    - ++rl_games.env_eval.image_size=224
+    - ++rl_games.env_eval.frameskip=1
+    - ++rl_games.env_eval.image_transform=raw_rgb
+    - ++rl_games.env_eval.prompt_mode=raw
+    - ++rl_games.env_eval.ghost_trail.history_frames=5
+    - ++rl_games.env_eval.ghost_trail.gamma=1.3
+    - ++rl_games.env_eval.ghost_trail.min_alpha=35
+    - ++rl_games.env_eval.ghost_trail.scroll_px_per_step=4.0
+    - ++rl_games.env_eval.ghost_trail.ground_fraction=0.22
+    - ++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.task_description=
+    - ++rl_games.env_eval.eval_parallel_envs=5
+    - ++rl_games.env_eval.enabled=true
+    - ++rl_games.env_eval.eval_backend=latency_bench
+    - ++rl_games.env_eval.distributed_mode=rank_sharded
+    - ++rl_games.env_eval.vectorized.enabled=false
+    - ++rl_games.env_eval.vectorized.batch_size=1
+    - ++rl_games.env_eval.latency.prompt_map_path=null
+    - ++rl_games.env_eval.latency.mode=single
+    - ++rl_games.env_eval.latency.values=[0]
+    - ++rl_games.env_eval.mid_train.enabled=true
+    - ++rl_games.env_eval.mid_train.interval_steps=250
+    - ++rl_games.env_eval.mid_train.latencies=[2]
+    - ++rl_games.env_eval.mid_train.num_episodes=20
+    - ++rl_games.env_eval.mid_train.max_steps_per_episode=3600
+    - ++rl_games.env_eval.post_train.enabled=true
+    - ++rl_games.env_eval.post_train.latencies=[2]
+    - ++rl_games.env_eval.post_train.num_episodes=50
+    - ++rl_games.env_eval.post_train.max_steps_per_episode=3600
+    - ++rl_games.task=flappy
+    - ++rl_games.initialization_mode=bridge
+    - ++rl_games.action_carrier=bridge
+    - ++bridge_base_model.repo_id.openvla=Qwen/Qwen3-VL-4B-Instruct
+    - ++bridge_base_model.repo_id.pi0=StarVLA/Qwen2.5-VL-3B-Instruct-Action
+    - ++bridge_base_model.repo_id.pi05=Qwen/Qwen3-VL-4B-Instruct
+    - ++bridge_base_model.repo_id.gr00t=Qwen/Qwen3-VL-4B-Instruct
+    - ++bridge_base_model.local_dir.openvla=playground/Pretrained_models/Qwen3-VL-4B-Instruct
+    - ++bridge_base_model.local_dir.pi0=playground/Pretrained_models/Qwen2.5-VL-3B-Instruct-Action
+    - ++bridge_base_model.local_dir.pi05=playground/Pretrained_models/Qwen3-VL-4B-Instruct
+    - ++bridge_base_model.local_dir.gr00t=playground/Pretrained_models/Qwen3-VL-4B-Instruct
+    - ++checkpoint.load=none
+    - ++checkpoint.hf_repo_id=null
+    - ++checkpoint.save_best_model=false
+    - ++checkpoint.save_final_model=true
+    - ++checkpoint.save_pt_file=false
+    - ++checkpoint.save_safetensors_file=true
+    - ++checkpoint.local.keep_last_n=1
+    - ++checkpoint.sync.enabled=false
+    - ++checkpoint.sync.repo_id=null
+    - ++checkpoint.sync.keep_last_n=0
+    - ++checkpoint.sync.sync_every_n_checkpoints=1
+    - ++checkpoint.sync.resume_policy=local_latest
+    - ++run_id=flappy_fix_latency_2_200ep_7k2steps_plain_multi
+    - ++output_dir=null
+    - ++config_yaml=null
+    - ++is_debug=false
+    - ++version_id=0.21
+    - ++run_root_dir=/workspace/latency-sensitive-bench/starVLA/results/Checkpoints
+    - trainer.is_resume=false
+    - trainer.pretrained_checkpoint=/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/checkpoints/_initialization/StarVLA_Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt
+    - trainer.resume_step=0
+    - ++datasets.vla_data.data_root_dir=/workspace/latency-sensitive-bench/starVLA/data/flappy_fix_latency_2_200ep_7k2steps
+    - ++datasets.vla_data.data_mix=flappy_train__bridge
+    - ++datasets.vla_data.eval_data_mix=flappy_train__bridge__val
+    - ++framework.qwenvl.base_vlm=/workspace/latency-sensitive-bench/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct
+    - ++rl_games.env_eval.latency.prompt_map_path=/workspace/latency-sensitive-bench/starVLA/data/flappy_fix_latency_2_200ep_7k2steps/flappy_train__bridge/latency_prompt_map.json
+  job:
+    name: train_starvla_hydra
+    chdir: false
+    override_dirname: ++auth.env_file=null,++auth.hf_token_env=HF_TOKEN,++auth.wandb_api_key_env=WANDB_API_KEY,++base_model.repo_id=Qwen/Qwen3-VL-4B-Instruct,++bridge_base_model.local_dir.gr00t=playground/Pretrained_models/Qwen3-VL-4B-Instruct,++bridge_base_model.local_dir.openvla=playground/Pretrained_models/Qwen3-VL-4B-Instruct,++bridge_base_model.local_dir.pi05=playground/Pretrained_models/Qwen3-VL-4B-Instruct,++bridge_base_model.local_dir.pi0=playground/Pretrained_models/Qwen2.5-VL-3B-Instruct-Action,++bridge_base_model.repo_id.gr00t=Qwen/Qwen3-VL-4B-Instruct,++bridge_base_model.repo_id.openvla=Qwen/Qwen3-VL-4B-Instruct,++bridge_base_model.repo_id.pi05=Qwen/Qwen3-VL-4B-Instruct,++bridge_base_model.repo_id.pi0=StarVLA/Qwen2.5-VL-3B-Instruct-Action,++checkpoint.hf_repo_id=null,++checkpoint.load=none,++checkpoint.local.keep_last_n=1,++checkpoint.save_best_model=false,++checkpoint.save_final_model=true,++checkpoint.save_pt_file=false,++checkpoint.save_safetensors_file=true,++checkpoint.sync.enabled=false,++checkpoint.sync.keep_last_n=0,++checkpoint.sync.repo_id=null,++checkpoint.sync.resume_policy=local_latest,++checkpoint.sync.sync_every_n_checkpoints=1,++config_yaml=null,++dataset.config_name=null,++dataset.converted_name=flappy_train,++dataset.debug_subset.enabled=false,++dataset.debug_subset.max_episodes=5,++dataset.debug_subset.suffix=debug,++dataset.episodes_per_latency=null,++dataset.force_download=false,++dataset.latency_filter=null,++dataset.max_episodes=null,++dataset.mixed_converted_name=flappy_mixed_latency_train,++dataset.mixed_latency_filter=null,++dataset.mixed_source_hf=,++dataset.setup_force=false,++dataset.single_converted_name=flappy_train,++dataset.single_latency_filter=null,++dataset.single_source_hf=,++dataset.skip_verification=false,++dataset.source_hf=,++dataset.source_subdir=null,++dataset.verify_rows=200,++datasets.vla_data.action_balance.action_key=action_id,++datasets.vla_data.action_balance.enabled=false,++datasets.vla_data.action_balance.flap_id=1,++datasets.vla_data.action_balance.noop_id=0,++datasets.vla_data.action_balance.strategy=balanced_epoch,++datasets.vla_data.action_balance.target_flap_fraction=0.3,++datasets.vla_data.action_type=discrete,++datasets.vla_data.custom_mixtures_path=null,++datasets.vla_data.data_mix=flappy_train,++datasets.vla_data.data_mix=flappy_train__bridge,++datasets.vla_data.data_root_dir=/workspace/latency-sensitive-bench/starVLA/data/flappy_fix_latency_2_200ep_7k2steps,++datasets.vla_data.data_root_dir=playground/Datasets/rl_games,++datasets.vla_data.dataset_py=lerobot_datasets,++datasets.vla_data.eval_data_mix=flappy_train__bridge__val,++datasets.vla_data.eval_data_mix=null,++datasets.vla_data.eval_num_workers=8,++datasets.vla_data.eval_sequential_step_sampling=null,++datasets.vla_data.image_mode=multiframe,++datasets.vla_data.include_state=true,++datasets.vla_data.latency_curriculum.computed_plan=null,++datasets.vla_data.latency_curriculum.enabled=false,++datasets.vla_data.latency_curriculum.eval_at_phase_end=false,++datasets.vla_data.latency_curriculum.final_equalization=true,++datasets.vla_data.latency_curriculum.latencies=null,++datasets.vla_data.latency_curriculum.new_latency_passes=1.0,++datasets.vla_data.latency_curriculum.phase_distributions=null,++datasets.vla_data.latency_curriculum.phase_steps=null,++datasets.vla_data.latency_curriculum.replay_passes=0.25,++datasets.vla_data.latency_curriculum.save_at_phase_end=false,++datasets.vla_data.latency_curriculum.step_budget_mode=auto,++datasets.vla_data.latency_curriculum.strategy=exclusive,++datasets.vla_data.latency_curriculum.target_total_passes=2.0,++datasets.vla_data.load_all_data_for_training=true,++datasets.vla_data.num_obs_frames=4,++datasets.vla_data.num_workers=8,++datasets.vla_data.obs_image_size=null,++datasets.vla_data.per_device_batch_size=16,++datasets.vla_data.persistent_workers=true,++datasets.vla_data.pin_memory=true,++datasets.vla_data.prefetch_factor=4,++datasets.vla_data.prompt_mode=raw,++datasets.vla_data.sequential_step_sampling=false,++datasets.vla_data.shuffle=true,++datasets.vla_data.stitch_grid=[2,2],++datasets.vla_data.video_backend=torchvision_av,++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.kv_memory.enabled=false,++framework.kv_memory.packed_train=false,++framework.kv_memory.rebased_sink=true,++framework.kv_memory.rollout_len=8,++framework.kv_memory.window=4,++framework.name=QwenOFT,++framework.qwenvl.attn_implementation=flash_attention_2,++framework.qwenvl.base_vlm=/workspace/latency-sensitive-bench/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct,++framework.qwenvl.base_vlm=playground/Pretrained_models/Qwen3-VL-4B-Instruct,++framework.qwenvl.enable_gradient_checkpointing=true,++framework.qwenvl.flex_backend=triton,++initialization.checkpoint_filename=checkpoints/steps_5000_pytorch_model.pt,++initialization.checkpoint_hf_repo_id=StarVLA/Qwen3VL-OFT-Bridge-RT-1,++initialization.checkpoint_local_dir=playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1,++is_debug=false,++output_dir=null,++paths.accelerate_config=starVLA/config/deepseeds/deepspeed_zero2.yaml,++paths.base_model_dir=playground/Pretrained_models/Qwen3-VL-4B-Instruct,++paths.dataset_cache_dir=null,++paths.dataset_local_dir=data/flappy_fix_latency_2_200ep_7k2steps,++paths.run_root_dir=results/Checkpoints,++rl_games.action_carrier=bridge,++rl_games.env_eval.distributed_mode=rank_sharded,++rl_games.env_eval.enabled=true,++rl_games.env_eval.eval_backend=latency_bench,++rl_games.env_eval.eval_parallel_envs=5,++rl_games.env_eval.fixed_episode_seeds=true,++rl_games.env_eval.frameskip=1,++rl_games.env_eval.ghost_trail.gamma=1.3,++rl_games.env_eval.ghost_trail.ground_fraction=0.22,++rl_games.env_eval.ghost_trail.history_frames=5,++rl_games.env_eval.ghost_trail.min_alpha=35,++rl_games.env_eval.ghost_trail.scroll_px_per_step=4.0,++rl_games.env_eval.image_size=224,++rl_games.env_eval.image_transform=raw_rgb,++rl_games.env_eval.latency.mode=single,++rl_games.env_eval.latency.prompt_map_path=/workspace/latency-sensitive-bench/starVLA/data/flappy_fix_latency_2_200ep_7k2steps/flappy_train__bridge/latency_prompt_map.json,++rl_games.env_eval.latency.prompt_map_path=null,++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=250,++rl_games.env_eval.mid_train.latencies=[2],++rl_games.env_eval.mid_train.max_steps_per_episode=3600,++rl_games.env_eval.mid_train.num_episodes=20,++rl_games.env_eval.post_train.enabled=true,++rl_games.env_eval.post_train.latencies=[2],++rl_games.env_eval.post_train.max_steps_per_episode=3600,++rl_games.env_eval.post_train.num_episodes=50,++rl_games.env_eval.prompt_mode=raw,++rl_games.env_eval.seed=42,++rl_games.env_eval.task_description=,++rl_games.env_eval.task_seed_stride=0,++rl_games.env_eval.vectorized.batch_size=1,++rl_games.env_eval.vectorized.enabled=false,++rl_games.initialization_mode=bridge,++rl_games.model_alias=openvla,++rl_games.task=flappy,++run_id=flappy_fix_latency_2_200ep_7k2steps_plain_multi,++run_root_dir=/workspace/latency-sensitive-bench/starVLA/results/Checkpoints,++run_root_dir=results/Checkpoints,++seed=42,++version_id=0.21,++wandb_entity=saberrr-zju,++wandb_project=starVLA_rl_games,++workspace_dir=WORKSPACE_DIR,env=flappy,init=bridge,mode=single,model=openvla,trainer.cc_f1_tolerance=1,trainer.distributed_backend=deepspeed,trainer.eval_action_classification=false,trainer.eval_action_classification_interval=null,trainer.eval_interval=250,trainer.eval_num_batches=50,trainer.freeze_llm_layers=[],trainer.freeze_modules=,trainer.freeze_tied_embedding=false,trainer.freeze_vit=false,trainer.gradient_accumulation_steps=4,trainer.gradient_clipping=1.0,trainer.is_resume=false,trainer.is_resume=false,trainer.learning_rate.action_model=0.0001,trainer.learning_rate.base=2e-05,trainer.learning_rate.qwen_vl_interface=1e-05,trainer.logging_frequency=1,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=4000,trainer.num_warmup_steps=100,trainer.optimizer.betas=[0.9,0.95],trainer.optimizer.eps=1e-08,trainer.optimizer.fused=true,trainer.optimizer.name=AdamW,trainer.optimizer.weight_decay=1e-08,trainer.per_latency_eval_num_batches=null,trainer.pretrained_checkpoint=/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/checkpoints/_initialization/StarVLA_Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt,trainer.pretrained_checkpoint=null,trainer.profile_timing.enabled=false,trainer.profile_timing.log_interval=10,trainer.reload_modules=null,trainer.resume_step=0,trainer.resume_step=0,trainer.save_format=pt,trainer.save_interval=500,trainer.scheduler_specific_kwargs.min_lr=1e-06,trainer.weight_decay=0.0
+    id: ???
+    num: ???
+    config_name: train
+    env_set: {}
+    env_copy: []
+    config:
+      override_dirname:
+        kv_sep: '='
+        item_sep: ','
+        exclude_keys: []
+  runtime:
+    version: 1.3.3
+    version_base: '1.1'
+    cwd: /workspace/latency-sensitive-bench/starVLA
+    config_sources:
+    - path: hydra.conf
+      schema: pkg
+      provider: hydra
+    - path: /workspace/latency-sensitive-bench/starVLA/examples/rl_games/config
+      schema: file
+      provider: main
+    - path: ''
+      schema: structured
+      provider: schema
+    output_dir: /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/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

flappy_fix_latency_2_200ep_7k2steps_plain_multi/hydra/.hydra/overrides.yaml

@@ -0,0 +1,208 @@
+- model=openvla
+- env=flappy
+- init=bridge
+- mode=single
+- ++framework.qwenvl.base_vlm=playground/Pretrained_models/Qwen3-VL-4B-Instruct
+- ++framework.qwenvl.attn_implementation=flash_attention_2
+- ++framework.qwenvl.flex_backend=triton
+- ++framework.qwenvl.enable_gradient_checkpointing=true
+- ++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
+- ++framework.action_model.action_dim=7
+- ++framework.action_model.action_env_dim=2
+- ++framework.kv_memory.enabled=false
+- ++framework.kv_memory.window=4
+- ++framework.kv_memory.rollout_len=8
+- ++framework.kv_memory.packed_train=false
+- ++framework.kv_memory.rebased_sink=true
+- ++framework.name=QwenOFT
+- ++datasets.vla_data.dataset_py=lerobot_datasets
+- ++datasets.vla_data.include_state=true
+- ++datasets.vla_data.data_root_dir=playground/Datasets/rl_games
+- ++datasets.vla_data.data_mix=flappy_train
+- ++datasets.vla_data.eval_data_mix=null
+- ++datasets.vla_data.custom_mixtures_path=null
+- ++datasets.vla_data.action_type=discrete
+- ++datasets.vla_data.sequential_step_sampling=false
+- ++datasets.vla_data.eval_sequential_step_sampling=null
+- ++datasets.vla_data.num_workers=8
+- ++datasets.vla_data.eval_num_workers=8
+- ++datasets.vla_data.prefetch_factor=4
+- ++datasets.vla_data.persistent_workers=true
+- ++datasets.vla_data.pin_memory=true
+- ++datasets.vla_data.shuffle=true
+- ++datasets.vla_data.action_balance.enabled=false
+- ++datasets.vla_data.action_balance.strategy=balanced_epoch
+- ++datasets.vla_data.action_balance.action_key=action_id
+- ++datasets.vla_data.action_balance.target_flap_fraction=0.3
+- ++datasets.vla_data.action_balance.noop_id=0
+- ++datasets.vla_data.action_balance.flap_id=1
+- ++datasets.vla_data.latency_curriculum.enabled=false
+- ++datasets.vla_data.latency_curriculum.strategy=exclusive
+- ++datasets.vla_data.latency_curriculum.latencies=null
+- ++datasets.vla_data.latency_curriculum.phase_steps=null
+- ++datasets.vla_data.latency_curriculum.phase_distributions=null
+- ++datasets.vla_data.latency_curriculum.new_latency_passes=1.0
+- ++datasets.vla_data.latency_curriculum.replay_passes=0.25
+- ++datasets.vla_data.latency_curriculum.target_total_passes=2.0
+- ++datasets.vla_data.latency_curriculum.final_equalization=true
+- ++datasets.vla_data.latency_curriculum.step_budget_mode=auto
+- ++datasets.vla_data.latency_curriculum.eval_at_phase_end=false
+- ++datasets.vla_data.latency_curriculum.save_at_phase_end=false
+- ++datasets.vla_data.latency_curriculum.computed_plan=null
+- ++datasets.vla_data.per_device_batch_size=16
+- ++datasets.vla_data.load_all_data_for_training=true
+- ++datasets.vla_data.num_obs_frames=4
+- ++datasets.vla_data.image_mode=multiframe
+- ++datasets.vla_data.prompt_mode=raw
+- ++datasets.vla_data.stitch_grid=[2,2]
+- ++datasets.vla_data.obs_image_size=null
+- ++datasets.vla_data.video_backend=torchvision_av
+- ++dataset.source_hf=
+- ++dataset.config_name=null
+- ++dataset.source_subdir=null
+- ++dataset.converted_name=flappy_train
+- ++dataset.single_source_hf=
+- ++dataset.mixed_source_hf=
+- ++dataset.single_converted_name=flappy_train
+- ++dataset.mixed_converted_name=flappy_mixed_latency_train
+- ++dataset.single_latency_filter=null
+- ++dataset.mixed_latency_filter=null
+- ++dataset.force_download=false
+- ++dataset.setup_force=false
+- ++dataset.skip_verification=false
+- ++dataset.verify_rows=200
+- ++dataset.max_episodes=null
+- ++dataset.episodes_per_latency=null
+- ++dataset.latency_filter=null
+- ++dataset.debug_subset.enabled=false
+- ++dataset.debug_subset.max_episodes=5
+- ++dataset.debug_subset.suffix=debug
+- ++base_model.repo_id=Qwen/Qwen3-VL-4B-Instruct
+- ++initialization.checkpoint_local_dir=playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1
+- ++initialization.checkpoint_hf_repo_id=StarVLA/Qwen3VL-OFT-Bridge-RT-1
+- ++initialization.checkpoint_filename=checkpoints/steps_5000_pytorch_model.pt
+- trainer.max_train_steps=4000
+- trainer.num_warmup_steps=100
+- trainer.save_interval=500
+- trainer.eval_interval=250
+- trainer.eval_num_batches=50
+- trainer.per_latency_eval_num_batches=null
+- trainer.eval_action_classification=false
+- trainer.eval_action_classification_interval=null
+- trainer.cc_f1_tolerance=1
+- 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.freeze_modules=
+- trainer.freeze_vit=false
+- trainer.freeze_tied_embedding=false
+- trainer.freeze_llm_layers=[]
+- trainer.loss_scale.vla=1.0
+- trainer.loss_scale.vlm=0.1
+- trainer.max_grad_norm=1.0
+- trainer.weight_decay=0.0
+- trainer.logging_frequency=1
+- trainer.profile_timing.enabled=false
+- trainer.profile_timing.log_interval=10
+- trainer.gradient_clipping=1.0
+- trainer.gradient_accumulation_steps=4
+- trainer.distributed_backend=deepspeed
+- trainer.is_resume=false
+- trainer.pretrained_checkpoint=null
+- trainer.resume_step=0
+- trainer.reload_modules=null
+- 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=true
+- trainer.save_format=pt
+- ++workspace_dir=WORKSPACE_DIR
+- ++run_root_dir=results/Checkpoints
+- ++seed=42
+- ++wandb_entity=saberrr-zju
+- ++wandb_project=starVLA_rl_games
+- ++auth.env_file=null
+- ++auth.hf_token_env=HF_TOKEN
+- ++auth.wandb_api_key_env=WANDB_API_KEY
+- ++paths.run_root_dir=results/Checkpoints
+- ++paths.dataset_local_dir=data/flappy_fix_latency_2_200ep_7k2steps
+- ++paths.dataset_cache_dir=null
+- ++paths.base_model_dir=playground/Pretrained_models/Qwen3-VL-4B-Instruct
+- ++paths.accelerate_config=starVLA/config/deepseeds/deepspeed_zero2.yaml
+- ++rl_games.model_alias=openvla
+- ++rl_games.env_eval.image_size=224
+- ++rl_games.env_eval.frameskip=1
+- ++rl_games.env_eval.image_transform=raw_rgb
+- ++rl_games.env_eval.prompt_mode=raw
+- ++rl_games.env_eval.ghost_trail.history_frames=5
+- ++rl_games.env_eval.ghost_trail.gamma=1.3
+- ++rl_games.env_eval.ghost_trail.min_alpha=35
+- ++rl_games.env_eval.ghost_trail.scroll_px_per_step=4.0
+- ++rl_games.env_eval.ghost_trail.ground_fraction=0.22
+- ++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.task_description=
+- ++rl_games.env_eval.eval_parallel_envs=5
+- ++rl_games.env_eval.enabled=true
+- ++rl_games.env_eval.eval_backend=latency_bench
+- ++rl_games.env_eval.distributed_mode=rank_sharded
+- ++rl_games.env_eval.vectorized.enabled=false
+- ++rl_games.env_eval.vectorized.batch_size=1
+- ++rl_games.env_eval.latency.prompt_map_path=null
+- ++rl_games.env_eval.latency.mode=single
+- ++rl_games.env_eval.latency.values=[0]
+- ++rl_games.env_eval.mid_train.enabled=true
+- ++rl_games.env_eval.mid_train.interval_steps=250
+- ++rl_games.env_eval.mid_train.latencies=[2]
+- ++rl_games.env_eval.mid_train.num_episodes=20
+- ++rl_games.env_eval.mid_train.max_steps_per_episode=3600
+- ++rl_games.env_eval.post_train.enabled=true
+- ++rl_games.env_eval.post_train.latencies=[2]
+- ++rl_games.env_eval.post_train.num_episodes=50
+- ++rl_games.env_eval.post_train.max_steps_per_episode=3600
+- ++rl_games.task=flappy
+- ++rl_games.initialization_mode=bridge
+- ++rl_games.action_carrier=bridge
+- ++bridge_base_model.repo_id.openvla=Qwen/Qwen3-VL-4B-Instruct
+- ++bridge_base_model.repo_id.pi0=StarVLA/Qwen2.5-VL-3B-Instruct-Action
+- ++bridge_base_model.repo_id.pi05=Qwen/Qwen3-VL-4B-Instruct
+- ++bridge_base_model.repo_id.gr00t=Qwen/Qwen3-VL-4B-Instruct
+- ++bridge_base_model.local_dir.openvla=playground/Pretrained_models/Qwen3-VL-4B-Instruct
+- ++bridge_base_model.local_dir.pi0=playground/Pretrained_models/Qwen2.5-VL-3B-Instruct-Action
+- ++bridge_base_model.local_dir.pi05=playground/Pretrained_models/Qwen3-VL-4B-Instruct
+- ++bridge_base_model.local_dir.gr00t=playground/Pretrained_models/Qwen3-VL-4B-Instruct
+- ++checkpoint.load=none
+- ++checkpoint.hf_repo_id=null
+- ++checkpoint.save_best_model=false
+- ++checkpoint.save_final_model=true
+- ++checkpoint.save_pt_file=false
+- ++checkpoint.save_safetensors_file=true
+- ++checkpoint.local.keep_last_n=1
+- ++checkpoint.sync.enabled=false
+- ++checkpoint.sync.repo_id=null
+- ++checkpoint.sync.keep_last_n=0
+- ++checkpoint.sync.sync_every_n_checkpoints=1
+- ++checkpoint.sync.resume_policy=local_latest
+- ++run_id=flappy_fix_latency_2_200ep_7k2steps_plain_multi
+- ++output_dir=null
+- ++config_yaml=null
+- ++is_debug=false
+- ++version_id=0.21
+- ++run_root_dir=/workspace/latency-sensitive-bench/starVLA/results/Checkpoints
+- trainer.is_resume=false
+- trainer.pretrained_checkpoint=/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/checkpoints/_initialization/StarVLA_Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt
+- trainer.resume_step=0
+- ++datasets.vla_data.data_root_dir=/workspace/latency-sensitive-bench/starVLA/data/flappy_fix_latency_2_200ep_7k2steps
+- ++datasets.vla_data.data_mix=flappy_train__bridge
+- ++datasets.vla_data.eval_data_mix=flappy_train__bridge__val
+- ++framework.qwenvl.base_vlm=/workspace/latency-sensitive-bench/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct
+- ++rl_games.env_eval.latency.prompt_map_path=/workspace/latency-sensitive-bench/starVLA/data/flappy_fix_latency_2_200ep_7k2steps/flappy_train__bridge/latency_prompt_map.json

flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1000/_progress/rank_0.json

@@ -0,0 +1 @@
+{"completed": 10}

flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1000/_progress/rank_1.json

@@ -0,0 +1 @@
+{"completed": 10}

flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1000/flappy/latency_2/episode_metrics.jsonl

@@ -0,0 +1,10 @@
+{"episode_id": 1, "episode_return": 6.200000092387199, "episode_return_env": 6.200000092387199, "game_score": 1.0, "mean_latency_ms": 66.66666666666667, "metadata": {"checkpoint_path": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1000/flappy/latency_2", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 43, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1000/flappy/latency_2", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "sample_index": null, "target_device": null, "trace_index": null}, "num_actions": 64, "num_dropped_actions": 0, "num_invalid_actions": 0, "p90_latency_ms": 66.66666666666667, "p99_latency_ms": 66.66666666666667, "return_raw": null, "survival_steps": 64}
+{"episode_id": 3, "episode_return": 3.90000007301569, "episode_return_env": 3.90000007301569, "game_score": 0.0, "mean_latency_ms": 66.66666666666667, "metadata": {"checkpoint_path": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1000/flappy/latency_2", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 45, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1000/flappy/latency_2", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "sample_index": null, "target_device": null, "trace_index": null}, "num_actions": 50, "num_dropped_actions": 0, "num_invalid_actions": 0, "p90_latency_ms": 66.66666666666667, "p99_latency_ms": 66.66666666666667, "return_raw": null, "survival_steps": 50}
+{"episode_id": 5, "episode_return": 5.000000089406967, "episode_return_env": 5.000000089406967, "game_score": 1.0, "mean_latency_ms": 66.66666666666667, "metadata": {"checkpoint_path": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1000/flappy/latency_2", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 47, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1000/flappy/latency_2", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "sample_index": null, "target_device": null, "trace_index": null}, "num_actions": 61, "num_dropped_actions": 0, "num_invalid_actions": 0, "p90_latency_ms": 66.66666666666667, "p99_latency_ms": 66.66666666666667, "return_raw": null, "survival_steps": 61}
+{"episode_id": 7, "episode_return": 5.000000089406967, "episode_return_env": 5.000000089406967, "game_score": 1.0, "mean_latency_ms": 66.66666666666667, "metadata": {"checkpoint_path": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1000/flappy/latency_2", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 49, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1000/flappy/latency_2", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "sample_index": null, "target_device": null, "trace_index": null}, "num_actions": 61, "num_dropped_actions": 0, "num_invalid_actions": 0, "p90_latency_ms": 66.66666666666667, "p99_latency_ms": 66.66666666666667, "return_raw": null, "survival_steps": 61}
+{"episode_id": 9, "episode_return": 5.000000089406967, "episode_return_env": 5.000000089406967, "game_score": 1.0, "mean_latency_ms": 66.66666666666667, "metadata": {"checkpoint_path": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1000/flappy/latency_2", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 51, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1000/flappy/latency_2", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "sample_index": null, "target_device": null, "trace_index": null}, "num_actions": 61, "num_dropped_actions": 0, "num_invalid_actions": 0, "p90_latency_ms": 66.66666666666667, "p99_latency_ms": 66.66666666666667, "return_raw": null, "survival_steps": 61}
+{"episode_id": 11, "episode_return": 48.50000058859587, "episode_return_env": 48.50000058859587, "game_score": 10.0, "mean_latency_ms": 66.66666666666667, "metadata": {"checkpoint_path": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1000/flappy/latency_2", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 53, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1000/flappy/latency_2", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "sample_index": null, "target_device": null, "trace_index": null}, "num_actions": 406, "num_dropped_actions": 0, "num_invalid_actions": 0, "p90_latency_ms": 66.66666666666667, "p99_latency_ms": 66.66666666666667, "return_raw": null, "survival_steps": 406}
+{"episode_id": 13, "episode_return": 3.90000007301569, "episode_return_env": 3.90000007301569, "game_score": 0.0, "mean_latency_ms": 66.66666666666667, "metadata": {"checkpoint_path": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1000/flappy/latency_2", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 55, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1000/flappy/latency_2", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "sample_index": null, "target_device": null, "trace_index": null}, "num_actions": 50, "num_dropped_actions": 0, "num_invalid_actions": 0, "p90_latency_ms": 66.66666666666667, "p99_latency_ms": 66.66666666666667, "return_raw": null, "survival_steps": 50}
+{"episode_id": 15, "episode_return": 5.000000089406967, "episode_return_env": 5.000000089406967, "game_score": 1.0, "mean_latency_ms": 66.66666666666667, "metadata": {"checkpoint_path": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1000/flappy/latency_2", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 57, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1000/flappy/latency_2", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "sample_index": null, "target_device": null, "trace_index": null}, "num_actions": 61, "num_dropped_actions": 0, "num_invalid_actions": 0, "p90_latency_ms": 66.66666666666667, "p99_latency_ms": 66.66666666666667, "return_raw": null, "survival_steps": 61}
+{"episode_id": 17, "episode_return": 284.8000033646822, "episode_return_env": 284.8000033646822, "game_score": 60.0, "mean_latency_ms": 66.66666666666667, "metadata": {"checkpoint_path": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1000/flappy/latency_2", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 59, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1000/flappy/latency_2", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "sample_index": null, "target_device": null, "trace_index": null}, "num_actions": 2319, "num_dropped_actions": 0, "num_invalid_actions": 0, "p90_latency_ms": 66.66666666666667, "p99_latency_ms": 66.66666666666667, "return_raw": null, "survival_steps": 2319}
+{"episode_id": 19, "episode_return": 5.000000089406967, "episode_return_env": 5.000000089406967, "game_score": 1.0, "mean_latency_ms": 66.66666666666667, "metadata": {"checkpoint_path": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1000/flappy/latency_2", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 61, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1000/flappy/latency_2", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "sample_index": null, "target_device": null, "trace_index": null}, "num_actions": 61, "num_dropped_actions": 0, "num_invalid_actions": 0, "p90_latency_ms": 66.66666666666667, "p99_latency_ms": 66.66666666666667, "return_raw": null, "survival_steps": 61}

flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1250/_progress/rank_0.json

@@ -0,0 +1 @@
+{"completed": 10}

flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1250/_progress/rank_1.json

@@ -0,0 +1 @@
+{"completed": 10}

flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1250/flappy/latency_2/episode_metrics.jsonl

@@ -0,0 +1,10 @@
+{"episode_id": 0, "episode_return": 118.70000141113997, "episode_return_env": 118.70000141113997, "game_score": 25.0, "mean_latency_ms": 66.66666666666667, "metadata": {"checkpoint_path": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1250/flappy/latency_2", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 42, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1250/flappy/latency_2", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "sample_index": null, "target_device": null, "trace_index": null}, "num_actions": 973, "num_dropped_actions": 0, "num_invalid_actions": 0, "p90_latency_ms": 66.66666666666667, "p99_latency_ms": 66.66666666666667, "return_raw": null, "survival_steps": 973}
+{"episode_id": 2, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 66.66666666666667, "metadata": {"checkpoint_path": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1250/flappy/latency_2", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 44, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1250/flappy/latency_2", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "sample_index": null, "target_device": null, "trace_index": null}, "num_actions": 3600, "num_dropped_actions": 0, "num_invalid_actions": 0, "p90_latency_ms": 66.66666666666667, "p99_latency_ms": 66.66666666666667, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 4, "episode_return": 102.90000123530626, "episode_return_env": 102.90000123530626, "game_score": 21.0, "mean_latency_ms": 66.66666666666667, "metadata": {"checkpoint_path": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1250/flappy/latency_2", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 46, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1250/flappy/latency_2", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "sample_index": null, "target_device": null, "trace_index": null}, "num_actions": 851, "num_dropped_actions": 0, "num_invalid_actions": 0, "p90_latency_ms": 66.66666666666667, "p99_latency_ms": 66.66666666666667, "return_raw": null, "survival_steps": 851}
+{"episode_id": 6, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 66.66666666666667, "metadata": {"checkpoint_path": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1250/flappy/latency_2", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 48, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1250/flappy/latency_2", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "sample_index": null, "target_device": null, "trace_index": null}, "num_actions": 3600, "num_dropped_actions": 0, "num_invalid_actions": 0, "p90_latency_ms": 66.66666666666667, "p99_latency_ms": 66.66666666666667, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 8, "episode_return": 426.60000501573086, "episode_return_env": 426.60000501573086, "game_score": 91.0, "mean_latency_ms": 66.66666666666667, "metadata": {"checkpoint_path": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1250/flappy/latency_2", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 50, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1250/flappy/latency_2", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "sample_index": null, "target_device": null, "trace_index": null}, "num_actions": 3458, "num_dropped_actions": 0, "num_invalid_actions": 0, "p90_latency_ms": 66.66666666666667, "p99_latency_ms": 66.66666666666667, "return_raw": null, "survival_steps": 3458}
+{"episode_id": 10, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 66.66666666666667, "metadata": {"checkpoint_path": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1250/flappy/latency_2", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 52, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1250/flappy/latency_2", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "sample_index": null, "target_device": null, "trace_index": null}, "num_actions": 3600, "num_dropped_actions": 0, "num_invalid_actions": 0, "p90_latency_ms": 66.66666666666667, "p99_latency_ms": 66.66666666666667, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 12, "episode_return": 218.40000258386135, "episode_return_env": 218.40000258386135, "game_score": 46.0, "mean_latency_ms": 66.66666666666667, "metadata": {"checkpoint_path": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1250/flappy/latency_2", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 54, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1250/flappy/latency_2", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "sample_index": null, "target_device": null, "trace_index": null}, "num_actions": 1781, "num_dropped_actions": 0, "num_invalid_actions": 0, "p90_latency_ms": 66.66666666666667, "p99_latency_ms": 66.66666666666667, "return_raw": null, "survival_steps": 1781}
+{"episode_id": 14, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 66.66666666666667, "metadata": {"checkpoint_path": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1250/flappy/latency_2", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 56, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1250/flappy/latency_2", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "sample_index": null, "target_device": null, "trace_index": null}, "num_actions": 3600, "num_dropped_actions": 0, "num_invalid_actions": 0, "p90_latency_ms": 66.66666666666667, "p99_latency_ms": 66.66666666666667, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 16, "episode_return": 23.200000301003456, "episode_return_env": 23.200000301003456, "game_score": 4.0, "mean_latency_ms": 66.66666666666667, "metadata": {"checkpoint_path": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1250/flappy/latency_2", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 58, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1250/flappy/latency_2", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "sample_index": null, "target_device": null, "trace_index": null}, "num_actions": 207, "num_dropped_actions": 0, "num_invalid_actions": 0, "p90_latency_ms": 66.66666666666667, "p99_latency_ms": 66.66666666666667, "return_raw": null, "survival_steps": 207}
+{"episode_id": 18, "episode_return": 384.5000045225024, "episode_return_env": 384.5000045225024, "game_score": 82.0, "mean_latency_ms": 66.66666666666667, "metadata": {"checkpoint_path": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1250/flappy/latency_2", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 60, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1250/flappy/latency_2", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "sample_index": null, "target_device": null, "trace_index": null}, "num_actions": 3118, "num_dropped_actions": 0, "num_invalid_actions": 0, "p90_latency_ms": 66.66666666666667, "p99_latency_ms": 66.66666666666667, "return_raw": null, "survival_steps": 3118}

flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1500/_progress/rank_0.json

@@ -0,0 +1 @@
+{"completed": 10}

flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1500/_progress/rank_1.json

@@ -0,0 +1 @@
+{"completed": 10}

flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1500/flappy/latency_2/episode_metrics.jsonl

@@ -0,0 +1,10 @@
+{"episode_id": 1, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 66.66666666666667, "metadata": {"checkpoint_path": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1500/flappy/latency_2", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 43, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1500/flappy/latency_2", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "sample_index": null, "target_device": null, "trace_index": null}, "num_actions": 3600, "num_dropped_actions": 0, "num_invalid_actions": 0, "p90_latency_ms": 66.66666666666667, "p99_latency_ms": 66.66666666666667, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 3, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 66.66666666666667, "metadata": {"checkpoint_path": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1500/flappy/latency_2", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 45, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1500/flappy/latency_2", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "sample_index": null, "target_device": null, "trace_index": null}, "num_actions": 3600, "num_dropped_actions": 0, "num_invalid_actions": 0, "p90_latency_ms": 66.66666666666667, "p99_latency_ms": 66.66666666666667, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 5, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 66.66666666666667, "metadata": {"checkpoint_path": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1500/flappy/latency_2", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 47, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1500/flappy/latency_2", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "sample_index": null, "target_device": null, "trace_index": null}, "num_actions": 3600, "num_dropped_actions": 0, "num_invalid_actions": 0, "p90_latency_ms": 66.66666666666667, "p99_latency_ms": 66.66666666666667, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 7, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 66.66666666666667, "metadata": {"checkpoint_path": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1500/flappy/latency_2", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 49, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1500/flappy/latency_2", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "sample_index": null, "target_device": null, "trace_index": null}, "num_actions": 3600, "num_dropped_actions": 0, "num_invalid_actions": 0, "p90_latency_ms": 66.66666666666667, "p99_latency_ms": 66.66666666666667, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 9, "episode_return": 13.000000178813934, "episode_return_env": 13.000000178813934, "game_score": 2.0, "mean_latency_ms": 66.66666666666667, "metadata": {"checkpoint_path": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1500/flappy/latency_2", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 51, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1500/flappy/latency_2", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "sample_index": null, "target_device": null, "trace_index": null}, "num_actions": 123, "num_dropped_actions": 0, "num_invalid_actions": 0, "p90_latency_ms": 66.66666666666667, "p99_latency_ms": 66.66666666666667, "return_raw": null, "survival_steps": 123}
+{"episode_id": 11, "episode_return": 90.6000010818243, "episode_return_env": 90.6000010818243, "game_score": 19.0, "mean_latency_ms": 66.66666666666667, "metadata": {"checkpoint_path": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1500/flappy/latency_2", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 53, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1500/flappy/latency_2", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "sample_index": null, "target_device": null, "trace_index": null}, "num_actions": 746, "num_dropped_actions": 0, "num_invalid_actions": 0, "p90_latency_ms": 66.66666666666667, "p99_latency_ms": 66.66666666666667, "return_raw": null, "survival_steps": 746}
+{"episode_id": 13, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 66.66666666666667, "metadata": {"checkpoint_path": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1500/flappy/latency_2", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 55, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1500/flappy/latency_2", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "sample_index": null, "target_device": null, "trace_index": null}, "num_actions": 3600, "num_dropped_actions": 0, "num_invalid_actions": 0, "p90_latency_ms": 66.66666666666667, "p99_latency_ms": 66.66666666666667, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 15, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 66.66666666666667, "metadata": {"checkpoint_path": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1500/flappy/latency_2", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 57, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1500/flappy/latency_2", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "sample_index": null, "target_device": null, "trace_index": null}, "num_actions": 3600, "num_dropped_actions": 0, "num_invalid_actions": 0, "p90_latency_ms": 66.66666666666667, "p99_latency_ms": 66.66666666666667, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 17, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 66.66666666666667, "metadata": {"checkpoint_path": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1500/flappy/latency_2", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 59, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1500/flappy/latency_2", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "sample_index": null, "target_device": null, "trace_index": null}, "num_actions": 3600, "num_dropped_actions": 0, "num_invalid_actions": 0, "p90_latency_ms": 66.66666666666667, "p99_latency_ms": 66.66666666666667, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 19, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 66.66666666666667, "metadata": {"checkpoint_path": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1500/flappy/latency_2", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 61, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1500/flappy/latency_2", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "sample_index": null, "target_device": null, "trace_index": null}, "num_actions": 3600, "num_dropped_actions": 0, "num_invalid_actions": 0, "p90_latency_ms": 66.66666666666667, "p99_latency_ms": 66.66666666666667, "return_raw": null, "survival_steps": 3600}

flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1750/_progress/rank_0.json

@@ -0,0 +1 @@
+{"completed": 10}

flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1750/_progress/rank_1.json

@@ -0,0 +1 @@
+{"completed": 10}

flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1750/flappy/latency_2/episode_metrics.jsonl

@@ -0,0 +1,10 @@
+{"episode_id": 0, "episode_return": 379.5000044628978, "episode_return_env": 379.5000044628978, "game_score": 81.0, "mean_latency_ms": 66.66666666666667, "metadata": {"checkpoint_path": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1750/flappy/latency_2", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 42, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1750/flappy/latency_2", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "sample_index": null, "target_device": null, "trace_index": null}, "num_actions": 3077, "num_dropped_actions": 0, "num_invalid_actions": 0, "p90_latency_ms": 66.66666666666667, "p99_latency_ms": 66.66666666666667, "return_raw": null, "survival_steps": 3077}
+{"episode_id": 2, "episode_return": 132.60000157356262, "episode_return_env": 132.60000157356262, "game_score": 28.0, "mean_latency_ms": 66.66666666666667, "metadata": {"checkpoint_path": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1750/flappy/latency_2", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 44, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1750/flappy/latency_2", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "sample_index": null, "target_device": null, "trace_index": null}, "num_actions": 1085, "num_dropped_actions": 0, "num_invalid_actions": 0, "p90_latency_ms": 66.66666666666667, "p99_latency_ms": 66.66666666666667, "return_raw": null, "survival_steps": 1085}
+{"episode_id": 4, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 66.66666666666667, "metadata": {"checkpoint_path": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1750/flappy/latency_2", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 46, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1750/flappy/latency_2", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "sample_index": null, "target_device": null, "trace_index": null}, "num_actions": 3600, "num_dropped_actions": 0, "num_invalid_actions": 0, "p90_latency_ms": 66.66666666666667, "p99_latency_ms": 66.66666666666667, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 6, "episode_return": 62.80000075697899, "episode_return_env": 62.80000075697899, "game_score": 13.0, "mean_latency_ms": 66.66666666666667, "metadata": {"checkpoint_path": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1750/flappy/latency_2", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 48, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1750/flappy/latency_2", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "sample_index": null, "target_device": null, "trace_index": null}, "num_actions": 522, "num_dropped_actions": 0, "num_invalid_actions": 0, "p90_latency_ms": 66.66666666666667, "p99_latency_ms": 66.66666666666667, "return_raw": null, "survival_steps": 522}
+{"episode_id": 8, "episode_return": 151.00000178813934, "episode_return_env": 151.00000178813934, "game_score": 32.0, "mean_latency_ms": 66.66666666666667, "metadata": {"checkpoint_path": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1750/flappy/latency_2", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 50, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1750/flappy/latency_2", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "sample_index": null, "target_device": null, "trace_index": null}, "num_actions": 1233, "num_dropped_actions": 0, "num_invalid_actions": 0, "p90_latency_ms": 66.66666666666667, "p99_latency_ms": 66.66666666666667, "return_raw": null, "survival_steps": 1233}
+{"episode_id": 10, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 66.66666666666667, "metadata": {"checkpoint_path": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1750/flappy/latency_2", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 52, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1750/flappy/latency_2", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "sample_index": null, "target_device": null, "trace_index": null}, "num_actions": 3600, "num_dropped_actions": 0, "num_invalid_actions": 0, "p90_latency_ms": 66.66666666666667, "p99_latency_ms": 66.66666666666667, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 12, "episode_return": 90.6000010818243, "episode_return_env": 90.6000010818243, "game_score": 19.0, "mean_latency_ms": 66.66666666666667, "metadata": {"checkpoint_path": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1750/flappy/latency_2", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 54, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1750/flappy/latency_2", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "sample_index": null, "target_device": null, "trace_index": null}, "num_actions": 746, "num_dropped_actions": 0, "num_invalid_actions": 0, "p90_latency_ms": 66.66666666666667, "p99_latency_ms": 66.66666666666667, "return_raw": null, "survival_steps": 746}
+{"episode_id": 14, "episode_return": 361.00000424683094, "episode_return_env": 361.00000424683094, "game_score": 77.0, "mean_latency_ms": 66.66666666666667, "metadata": {"checkpoint_path": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1750/flappy/latency_2", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 56, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1750/flappy/latency_2", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "sample_index": null, "target_device": null, "trace_index": null}, "num_actions": 2928, "num_dropped_actions": 0, "num_invalid_actions": 0, "p90_latency_ms": 66.66666666666667, "p99_latency_ms": 66.66666666666667, "return_raw": null, "survival_steps": 2928}
+{"episode_id": 16, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 66.66666666666667, "metadata": {"checkpoint_path": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1750/flappy/latency_2", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 58, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1750/flappy/latency_2", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "sample_index": null, "target_device": null, "trace_index": null}, "num_actions": 3600, "num_dropped_actions": 0, "num_invalid_actions": 0, "p90_latency_ms": 66.66666666666667, "p99_latency_ms": 66.66666666666667, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 18, "episode_return": 34.60000042617321, "episode_return_env": 34.60000042617321, "game_score": 7.0, "mean_latency_ms": 66.66666666666667, "metadata": {"checkpoint_path": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1750/flappy/latency_2", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 60, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1750/flappy/latency_2", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "sample_index": null, "target_device": null, "trace_index": null}, "num_actions": 294, "num_dropped_actions": 0, "num_invalid_actions": 0, "p90_latency_ms": 66.66666666666667, "p99_latency_ms": 66.66666666666667, "return_raw": null, "survival_steps": 294}

flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2000/_progress/rank_0.json

@@ -0,0 +1 @@
+{"completed": 10}

flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2000/_progress/rank_1.json

@@ -0,0 +1 @@
+{"completed": 10}

flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2000/flappy/latency_2/episode_metrics.jsonl

@@ -0,0 +1,10 @@
+{"episode_id": 0, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 66.66666666666667, "metadata": {"checkpoint_path": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2000/flappy/latency_2", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 42, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2000/flappy/latency_2", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "sample_index": null, "target_device": null, "trace_index": null}, "num_actions": 3600, "num_dropped_actions": 0, "num_invalid_actions": 0, "p90_latency_ms": 66.66666666666667, "p99_latency_ms": 66.66666666666667, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 2, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 66.66666666666667, "metadata": {"checkpoint_path": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2000/flappy/latency_2", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 44, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2000/flappy/latency_2", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "sample_index": null, "target_device": null, "trace_index": null}, "num_actions": 3600, "num_dropped_actions": 0, "num_invalid_actions": 0, "p90_latency_ms": 66.66666666666667, "p99_latency_ms": 66.66666666666667, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 4, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 66.66666666666667, "metadata": {"checkpoint_path": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2000/flappy/latency_2", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 46, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2000/flappy/latency_2", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "sample_index": null, "target_device": null, "trace_index": null}, "num_actions": 3600, "num_dropped_actions": 0, "num_invalid_actions": 0, "p90_latency_ms": 66.66666666666667, "p99_latency_ms": 66.66666666666667, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 6, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 66.66666666666667, "metadata": {"checkpoint_path": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2000/flappy/latency_2", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 48, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2000/flappy/latency_2", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "sample_index": null, "target_device": null, "trace_index": null}, "num_actions": 3600, "num_dropped_actions": 0, "num_invalid_actions": 0, "p90_latency_ms": 66.66666666666667, "p99_latency_ms": 66.66666666666667, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 8, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 66.66666666666667, "metadata": {"checkpoint_path": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2000/flappy/latency_2", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 50, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2000/flappy/latency_2", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "sample_index": null, "target_device": null, "trace_index": null}, "num_actions": 3600, "num_dropped_actions": 0, "num_invalid_actions": 0, "p90_latency_ms": 66.66666666666667, "p99_latency_ms": 66.66666666666667, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 10, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 66.66666666666667, "metadata": {"checkpoint_path": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2000/flappy/latency_2", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 52, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2000/flappy/latency_2", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "sample_index": null, "target_device": null, "trace_index": null}, "num_actions": 3600, "num_dropped_actions": 0, "num_invalid_actions": 0, "p90_latency_ms": 66.66666666666667, "p99_latency_ms": 66.66666666666667, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 12, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 66.66666666666667, "metadata": {"checkpoint_path": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2000/flappy/latency_2", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 54, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2000/flappy/latency_2", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "sample_index": null, "target_device": null, "trace_index": null}, "num_actions": 3600, "num_dropped_actions": 0, "num_invalid_actions": 0, "p90_latency_ms": 66.66666666666667, "p99_latency_ms": 66.66666666666667, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 14, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 66.66666666666667, "metadata": {"checkpoint_path": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2000/flappy/latency_2", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 56, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2000/flappy/latency_2", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "sample_index": null, "target_device": null, "trace_index": null}, "num_actions": 3600, "num_dropped_actions": 0, "num_invalid_actions": 0, "p90_latency_ms": 66.66666666666667, "p99_latency_ms": 66.66666666666667, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 16, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 66.66666666666667, "metadata": {"checkpoint_path": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2000/flappy/latency_2", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 58, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2000/flappy/latency_2", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "sample_index": null, "target_device": null, "trace_index": null}, "num_actions": 3600, "num_dropped_actions": 0, "num_invalid_actions": 0, "p90_latency_ms": 66.66666666666667, "p99_latency_ms": 66.66666666666667, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 18, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 66.66666666666667, "metadata": {"checkpoint_path": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2000/flappy/latency_2", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 60, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2000/flappy/latency_2", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "sample_index": null, "target_device": null, "trace_index": null}, "num_actions": 3600, "num_dropped_actions": 0, "num_invalid_actions": 0, "p90_latency_ms": 66.66666666666667, "p99_latency_ms": 66.66666666666667, "return_raw": null, "survival_steps": 3600}

flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2250/_progress/rank_0.json

@@ -0,0 +1 @@
+{"completed": 10}

flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2250/_progress/rank_1.json

@@ -0,0 +1 @@
+{"completed": 10}

flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2250/flappy/latency_2/episode_metrics.jsonl

@@ -0,0 +1,10 @@
+{"episode_id": 0, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 66.66666666666667, "metadata": {"checkpoint_path": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2250/flappy/latency_2", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 42, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2250/flappy/latency_2", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "sample_index": null, "target_device": null, "trace_index": null}, "num_actions": 3600, "num_dropped_actions": 0, "num_invalid_actions": 0, "p90_latency_ms": 66.66666666666667, "p99_latency_ms": 66.66666666666667, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 2, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 66.66666666666667, "metadata": {"checkpoint_path": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2250/flappy/latency_2", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 44, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2250/flappy/latency_2", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "sample_index": null, "target_device": null, "trace_index": null}, "num_actions": 3600, "num_dropped_actions": 0, "num_invalid_actions": 0, "p90_latency_ms": 66.66666666666667, "p99_latency_ms": 66.66666666666667, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 4, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 66.66666666666667, "metadata": {"checkpoint_path": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2250/flappy/latency_2", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 46, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2250/flappy/latency_2", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "sample_index": null, "target_device": null, "trace_index": null}, "num_actions": 3600, "num_dropped_actions": 0, "num_invalid_actions": 0, "p90_latency_ms": 66.66666666666667, "p99_latency_ms": 66.66666666666667, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 6, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 66.66666666666667, "metadata": {"checkpoint_path": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2250/flappy/latency_2", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 48, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2250/flappy/latency_2", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "sample_index": null, "target_device": null, "trace_index": null}, "num_actions": 3600, "num_dropped_actions": 0, "num_invalid_actions": 0, "p90_latency_ms": 66.66666666666667, "p99_latency_ms": 66.66666666666667, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 8, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 66.66666666666667, "metadata": {"checkpoint_path": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2250/flappy/latency_2", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 50, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2250/flappy/latency_2", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "sample_index": null, "target_device": null, "trace_index": null}, "num_actions": 3600, "num_dropped_actions": 0, "num_invalid_actions": 0, "p90_latency_ms": 66.66666666666667, "p99_latency_ms": 66.66666666666667, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 10, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 66.66666666666667, "metadata": {"checkpoint_path": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2250/flappy/latency_2", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 52, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2250/flappy/latency_2", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "sample_index": null, "target_device": null, "trace_index": null}, "num_actions": 3600, "num_dropped_actions": 0, "num_invalid_actions": 0, "p90_latency_ms": 66.66666666666667, "p99_latency_ms": 66.66666666666667, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 12, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 66.66666666666667, "metadata": {"checkpoint_path": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2250/flappy/latency_2", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 54, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2250/flappy/latency_2", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "sample_index": null, "target_device": null, "trace_index": null}, "num_actions": 3600, "num_dropped_actions": 0, "num_invalid_actions": 0, "p90_latency_ms": 66.66666666666667, "p99_latency_ms": 66.66666666666667, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 14, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 66.66666666666667, "metadata": {"checkpoint_path": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2250/flappy/latency_2", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 56, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2250/flappy/latency_2", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "sample_index": null, "target_device": null, "trace_index": null}, "num_actions": 3600, "num_dropped_actions": 0, "num_invalid_actions": 0, "p90_latency_ms": 66.66666666666667, "p99_latency_ms": 66.66666666666667, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 16, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 66.66666666666667, "metadata": {"checkpoint_path": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2250/flappy/latency_2", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 58, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2250/flappy/latency_2", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "sample_index": null, "target_device": null, "trace_index": null}, "num_actions": 3600, "num_dropped_actions": 0, "num_invalid_actions": 0, "p90_latency_ms": 66.66666666666667, "p99_latency_ms": 66.66666666666667, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 18, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 66.66666666666667, "metadata": {"checkpoint_path": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2250/flappy/latency_2", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 60, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2250/flappy/latency_2", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "sample_index": null, "target_device": null, "trace_index": null}, "num_actions": 3600, "num_dropped_actions": 0, "num_invalid_actions": 0, "p90_latency_ms": 66.66666666666667, "p99_latency_ms": 66.66666666666667, "return_raw": null, "survival_steps": 3600}

flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_250/_progress/rank_0.json

@@ -0,0 +1 @@
+{"completed": 10}

flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_250/_progress/rank_1.json

@@ -0,0 +1 @@
+{"completed": 10}

flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_250/flappy/latency_2/episode_metrics.jsonl

@@ -0,0 +1,10 @@
+{"episode_id": 0, "episode_return": 6.100000090897083, "episode_return_env": 6.100000090897083, "game_score": 1.0, "mean_latency_ms": 66.66666666666667, "metadata": {"checkpoint_path": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_250/flappy/latency_2", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 42, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_250/flappy/latency_2", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "sample_index": null, "target_device": null, "trace_index": null}, "num_actions": 63, "num_dropped_actions": 0, "num_invalid_actions": 0, "p90_latency_ms": 66.66666666666667, "p99_latency_ms": 66.66666666666667, "return_raw": null, "survival_steps": 63}
+{"episode_id": 2, "episode_return": 4.800000086426735, "episode_return_env": 4.800000086426735, "game_score": 0.0, "mean_latency_ms": 66.66666666666667, "metadata": {"checkpoint_path": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_250/flappy/latency_2", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 44, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_250/flappy/latency_2", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "sample_index": null, "target_device": null, "trace_index": null}, "num_actions": 59, "num_dropped_actions": 0, "num_invalid_actions": 0, "p90_latency_ms": 66.66666666666667, "p99_latency_ms": 66.66666666666667, "return_raw": null, "survival_steps": 59}
+{"episode_id": 4, "episode_return": 6.000000089406967, "episode_return_env": 6.000000089406967, "game_score": 1.0, "mean_latency_ms": 66.66666666666667, "metadata": {"checkpoint_path": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_250/flappy/latency_2", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 46, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_250/flappy/latency_2", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "sample_index": null, "target_device": null, "trace_index": null}, "num_actions": 62, "num_dropped_actions": 0, "num_invalid_actions": 0, "p90_latency_ms": 66.66666666666667, "p99_latency_ms": 66.66666666666667, "return_raw": null, "survival_steps": 62}
+{"episode_id": 6, "episode_return": 6.100000090897083, "episode_return_env": 6.100000090897083, "game_score": 1.0, "mean_latency_ms": 66.66666666666667, "metadata": {"checkpoint_path": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_250/flappy/latency_2", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 48, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_250/flappy/latency_2", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "sample_index": null, "target_device": null, "trace_index": null}, "num_actions": 63, "num_dropped_actions": 0, "num_invalid_actions": 0, "p90_latency_ms": 66.66666666666667, "p99_latency_ms": 66.66666666666667, "return_raw": null, "survival_steps": 63}
+{"episode_id": 8, "episode_return": 6.000000089406967, "episode_return_env": 6.000000089406967, "game_score": 1.0, "mean_latency_ms": 66.66666666666667, "metadata": {"checkpoint_path": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_250/flappy/latency_2", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 50, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_250/flappy/latency_2", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "sample_index": null, "target_device": null, "trace_index": null}, "num_actions": 62, "num_dropped_actions": 0, "num_invalid_actions": 0, "p90_latency_ms": 66.66666666666667, "p99_latency_ms": 66.66666666666667, "return_raw": null, "survival_steps": 62}
+{"episode_id": 10, "episode_return": 6.100000090897083, "episode_return_env": 6.100000090897083, "game_score": 1.0, "mean_latency_ms": 66.66666666666667, "metadata": {"checkpoint_path": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_250/flappy/latency_2", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 52, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_250/flappy/latency_2", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "sample_index": null, "target_device": null, "trace_index": null}, "num_actions": 63, "num_dropped_actions": 0, "num_invalid_actions": 0, "p90_latency_ms": 66.66666666666667, "p99_latency_ms": 66.66666666666667, "return_raw": null, "survival_steps": 63}
+{"episode_id": 12, "episode_return": 6.100000090897083, "episode_return_env": 6.100000090897083, "game_score": 1.0, "mean_latency_ms": 66.66666666666667, "metadata": {"checkpoint_path": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_250/flappy/latency_2", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 54, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_250/flappy/latency_2", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "sample_index": null, "target_device": null, "trace_index": null}, "num_actions": 63, "num_dropped_actions": 0, "num_invalid_actions": 0, "p90_latency_ms": 66.66666666666667, "p99_latency_ms": 66.66666666666667, "return_raw": null, "survival_steps": 63}
+{"episode_id": 14, "episode_return": 6.100000090897083, "episode_return_env": 6.100000090897083, "game_score": 1.0, "mean_latency_ms": 66.66666666666667, "metadata": {"checkpoint_path": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_250/flappy/latency_2", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 56, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_250/flappy/latency_2", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "sample_index": null, "target_device": null, "trace_index": null}, "num_actions": 63, "num_dropped_actions": 0, "num_invalid_actions": 0, "p90_latency_ms": 66.66666666666667, "p99_latency_ms": 66.66666666666667, "return_raw": null, "survival_steps": 63}
+{"episode_id": 16, "episode_return": 6.000000089406967, "episode_return_env": 6.000000089406967, "game_score": 1.0, "mean_latency_ms": 66.66666666666667, "metadata": {"checkpoint_path": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_250/flappy/latency_2", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 58, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_250/flappy/latency_2", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "sample_index": null, "target_device": null, "trace_index": null}, "num_actions": 62, "num_dropped_actions": 0, "num_invalid_actions": 0, "p90_latency_ms": 66.66666666666667, "p99_latency_ms": 66.66666666666667, "return_raw": null, "survival_steps": 62}
+{"episode_id": 18, "episode_return": 6.000000089406967, "episode_return_env": 6.000000089406967, "game_score": 1.0, "mean_latency_ms": 66.66666666666667, "metadata": {"checkpoint_path": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_250/flappy/latency_2", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 60, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_250/flappy/latency_2", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "sample_index": null, "target_device": null, "trace_index": null}, "num_actions": 62, "num_dropped_actions": 0, "num_invalid_actions": 0, "p90_latency_ms": 66.66666666666667, "p99_latency_ms": 66.66666666666667, "return_raw": null, "survival_steps": 62}

flappy_fix_latency_2_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2500/_progress/rank_0.json

@@ -0,0 +1 @@
+{"completed": 10}