Publish hist8 zero flappy qwenpi_v3 step5000

zero-latency/flappy/vla/starvla-qwenpi_v3-h1/flappy-qwenpi_v3-hist8-zero-scratch-s42-v2/README.md

@@ -0,0 +1,36 @@
+# flappy / qwenpi_v3 — hist8 zero training
+
+Run `flappy-qwenpi_v3-hist8-zero-scratch-s42-v2`; final optimizer step **5000**, seed **42**. [W&B](https://wandb.ai/saberrr-zju/starvla_tasks/runs/9be1cc91).
+
+Checkpoint SHA256: `f6bf2adb10f48fba4c66564bf278966aad004231ef354b10ffb1a6b7e513fc60`.
+
+## Model and inputs
+
+Full StarVLA state dict in `checkpoints/model.pt` (original evaluated bytes). Initialize the matching Qwen model with `config.yaml` and `dataset_statistics.json`. Base: `Qwen/Qwen3-VL-4B-Instruct@ebb281ec70b05090aa6165b016eac8ec08e71b17`; model-specific action head trained from random initialization.
+
+One current RGB image, eight prior causal decision histories (24 raw 0/1 values), action horizon 1. `h1` in this repository path is the action horizon; `hist8` is input history. Pass history as `transport`, preserve all-zero reset, and use identity state normalization. Prompt and action order are in `task_contract.json` and `manifest.json`.
+
+Training: 2 GPUs, bf16/ZeRO-2, batch 64 per rank, accumulation 1, global batch 128; 16 dataloader workers per rank and prefetch 4. Backbone LR 3e-6, action head 1e-4; other groups retain the recorded configuration. Warmup 100, cosine min_lr_rate 1/30, 5000 steps. Full historical training configuration and code snapshot are included.
+
+## Formal evaluation
+
+| Environment | Episodes | Return mean ± std | Mean raw frames |
+|---|---:|---:|---:|
+| profile | 100 | 5.5820 ± 1.9567 | 62.50 |
+| zero | 100 | 408.7170 ± 88.4673 | 3311.57 |
+
+Seeds 1,000,000–1,000,099; parallel 32; capacity 1; FPS/decision Hz 10/10; max 3600 raw frames. Zero-trained models have both zero and profile evaluations; profile-trained models have the matching profile evaluation. `profiles/profile.json` is the exact model-specific latency profile, with SHA in provenance.
+
+Each `evaluations/<condition>/` contains the evaluated config, 100 episode records, action audit including raw-score summaries, verification, provenance and evaluator code snapshot. Evaluation configs retain original machine paths as execution evidence; replace checkpoint, model-config, backbone, profile and output paths when replaying locally. Deadly Corridor requires 160×120 RGB with HUD, no crosshair, weapon enabled, decals/particles disabled before resizing.
+
+These are one-training-seed results. Architectures use different latency profiles. Zero/profile training also uses different teachers and collected data, so their score difference is not an isolated learning-rate ablation.
+
+## Download
+
+```python
+from huggingface_hub import snapshot_download
+prefix = "zero-latency/flappy/vla/starvla-qwenpi_v3-h1/flappy-qwenpi_v3-hist8-zero-scratch-s42-v2"
+repo = snapshot_download("latency-sensitive-bench/benchmark-models", allow_patterns=[prefix + "/**"])
+```
+
+Use a pinned repository commit for reproduction. Preserve the original checkpoint SHA above. Optimizer/RNG resume state remains in the source training archive; this publication contains the inference checkpoint.

zero-latency/flappy/vla/starvla-qwenpi_v3-h1/flappy-qwenpi_v3-hist8-zero-scratch-s42-v2/bundle.complete.json

@@ -0,0 +1,4 @@
+{
+  "checkpoint_sha256": "f6bf2adb10f48fba4c66564bf278966aad004231ef354b10ffb1a6b7e513fc60",
+  "training_step": 5000
+}

zero-latency/flappy/vla/starvla-qwenpi_v3-h1/flappy-qwenpi_v3-hist8-zero-scratch-s42-v2/checkpoint.sha256

@@ -0,0 +1 @@
+f6bf2adb10f48fba4c66564bf278966aad004231ef354b10ffb1a6b7e513fc60  checkpoints/model.pt

zero-latency/flappy/vla/starvla-qwenpi_v3-h1/flappy-qwenpi_v3-hist8-zero-scratch-s42-v2/checkpoints/model.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f6bf2adb10f48fba4c66564bf278966aad004231ef354b10ffb1a6b7e513fc60
+size 10922651805

zero-latency/flappy/vla/starvla-qwenpi_v3-h1/flappy-qwenpi_v3-hist8-zero-scratch-s42-v2/collection_quality_report.json

@@ -0,0 +1,30 @@
+{
+  "episodes": 200,
+  "train_episodes": 180,
+  "val_episodes": 20,
+  "train_labels": 647783,
+  "val_labels": 70375,
+  "raw_frames_quantiles": [
+    1975.0,
+    3600.0,
+    3600.0,
+    3600.0,
+    3600.0
+  ],
+  "return_quantiles": [
+    242.3000028654933,
+    444.6000052243471,
+    444.6000052243471,
+    444.6000052243471,
+    444.6000052243471
+  ],
+  "termination_reasons": {
+    "horizon": 198,
+    "terminated": 2
+  },
+  "action_id_counts": {
+    "0": 658426,
+    "1": 59732
+  },
+  "previous_decision_admitted_fraction": 1.0
+}

zero-latency/flappy/vla/starvla-qwenpi_v3-h1/flappy-qwenpi_v3-hist8-zero-scratch-s42-v2/config.full.yaml

@@ -0,0 +1,297 @@
+framework:
+  name: QwenPI_v3
+  qwenvl:
+    base_vlm: /home/ubuntu/.cache/huggingface/hub/models--Qwen--Qwen3-VL-4B-Instruct/snapshots/ebb281ec70b05090aa6165b016eac8ec08e71b17
+    attn_implementation: flash_attention_2
+    vl_hidden_dim: 2560
+    num_vl_layers: 36
+    enable_gradient_checkpointing: true
+  action_model:
+    action_model_type: LayerwiseFM
+    action_dim: 2
+    state_dim: 24
+    state_encoding: continuous_projector
+    task_objective: null
+    action_horizon: 1
+    repeated_diffusion_steps: 2
+    num_inference_timesteps: 4
+    add_pos_embed: true
+    max_seq_len: 1024
+    num_target_vision_tokens: 32
+    noise_beta_alpha: 1.5
+    noise_beta_beta: 1.0
+    noise_s: 0.999
+    num_timestep_buckets: 1000
+    diffusion_model_cfg:
+      action_dit_hidden_dim: 1024
+      dropout: 0.2
+      final_dropout: true
+      interleave_self_attention: true
+      norm_type: ada_norm
+      positional_embeddings: null
+      attention_head_dim: 64
+      output_dim: 1024
+      num_layers: 36
+      input_embedding_dim: 1024
+      cross_attention_dim: 1024
+      num_attention_heads: 16
+    loss_type: l1
+    future_action_window_size: 0
+    past_action_window_size: 0
+    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: /home/ubuntu/codex/hist8-vla-v2-20260923/outputs/tasks/decision_history_vla_hist8_v2/data_zero/flappy
+    data_mix: lerobot
+    eval_data_mix: lerobot__val
+    custom_mixtures_path: /home/ubuntu/codex/hist8-vla-v2-20260923/outputs/tasks/decision_history_vla_hist8_v2/data_zero/flappy/_generated_mixtures/lerobot.json
+    action_type: discrete
+    sequential_step_sampling: false
+    eval_sequential_step_sampling: null
+    num_workers: 16
+    eval_num_workers: 16
+    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: 64
+    load_all_data_for_training: true
+    num_obs_frames: 1
+    image_mode: single
+    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
+  target_latency_unit: observation_steps
+  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: null
+  checkpoint_hf_repo_id: null
+  checkpoint_filename: null
+trainer:
+  max_train_steps: 5000
+  num_warmup_steps: 100
+  save_interval: 500
+  eval_interval: 500
+  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: 3.0e-05
+    qwen_vl_interface: 3.0e-06
+    action_model: 0.0001
+  lr_scheduler_type: cosine_with_min_lr
+  scheduler_specific_kwargs:
+    min_lr_rate: 0.03333333333333333
+  freeze_modules: ''
+  freeze_vit: false
+  freeze_tied_embedding: false
+  freeze_llm_layers: []
+  loss_scale:
+    vla: 1.0
+    vlm: 0.0
+  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: 1
+  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
+  hist8_diagnostics:
+    steps:
+    - 0
+    - 500
+    - 1000
+    - 2500
+    - 5000
+    sample_count: 2048
+    noise_seeds:
+    - 1729
+    - 1730
+    - 1731
+    - 1732
+    sample_seed: 42
+    batch_size: 8
+    bootstrap_repeats: 4000
+workspace_dir: WORKSPACE_DIR
+run_root_dir: /home/ubuntu/codex/hist8-vla-v2-20260923/outputs/tasks/decision_history_vla_hist8_v2/train
+seed: 42
+wandb_entity: saberrr-zju
+wandb_project: starvla_tasks
+auth:
+  env_file: null
+  hf_token_env: HF_TOKEN
+  wandb_api_key_env: WANDB_API_KEY
+paths:
+  run_root_dir: results/Checkpoints
+  dataset_local_dir: playground/Datasets/rl_games
+  dataset_cache_dir: null
+  base_model_dir: /home/ubuntu/.cache/huggingface/hub/models--Qwen--Qwen3-VL-4B-Instruct/snapshots/ebb281ec70b05090aa6165b016eac8ec08e71b17
+  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: pi-0.5
+  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
+    action_chunk_execution:
+      enabled: false
+      chunk_size: null
+    enabled: false
+    eval_backend: latency_bench
+    distributed_mode: none
+    vectorized:
+      enabled: false
+      batch_size: 1
+    latency:
+      prompt_map_path: null
+      mode: single
+      values:
+      - 0
+    mid_train:
+      enabled: false
+      interval_steps: 100
+      latencies:
+      - 0
+      num_episodes: 5
+      max_steps_per_episode: 3600
+    post_train:
+      enabled: false
+      latencies:
+      - 0
+      - 1
+      - 2
+      - 3
+      - 4
+      num_episodes: 5
+      max_steps_per_episode: 3600
+  task: flappy
+  initialization_mode: scratch
+  action_carrier: native
+model: pi05
+env: flappy
+init: scratch
+mode: single
+checkpoint:
+  load: none
+  hf_repo_id: null
+  save_best_model: false
+  save_final_model: true
+  save_pt_file: true
+  save_training_state: true
+  save_safetensors_file: false
+  local:
+    keep_last_n: 2
+  sync:
+    enabled: false
+    repo_id: null
+    keep_last_n: 0
+    sync_every_n_checkpoints: 1
+    resume_policy: local_latest
+run_id: flappy-qwenpi_v3-hist8-zero-scratch-s42-v2
+output_dir: /home/ubuntu/codex/hist8-vla-v2-20260923/outputs/tasks/decision_history_vla_hist8_v2/train/flappy-qwenpi_v3-hist8-zero-scratch-s42-v2
+config_yaml: /home/ubuntu/codex/hist8-vla-v2-20260923/outputs/tasks/decision_history_vla_hist8_v2/train/flappy-qwenpi_v3-hist8-zero-scratch-s42-v2/train.yaml
+is_debug: false
+version_id: '0.21'
+wandb_name: flappy-qwenpi_v3-hist8-zero-scratch-s42-v2
+wandb_group: decision-history-vla-h8-v2
+wandb_tags:
+- flappy
+- qwenpi_v3
+- hist8-v2
+- history01
+- scratch
+- zero-latency
+training_latency_condition: zero-latency

zero-latency/flappy/vla/starvla-qwenpi_v3-h1/flappy-qwenpi_v3-hist8-zero-scratch-s42-v2/config.yaml

@@ -0,0 +1,57 @@
+framework:
+  name: QwenPI_v3
+  qwenvl:
+    base_vlm: Qwen/Qwen3-VL-4B-Instruct
+    attn_implementation: flash_attention_2
+    vl_hidden_dim: 2560
+    num_vl_layers: 36
+    enable_gradient_checkpointing: true
+  action_model:
+    action_model_type: LayerwiseFM
+    action_dim: 2
+    state_dim: 24
+    state_encoding: continuous_projector
+    task_objective: null
+    action_horizon: 1
+    repeated_diffusion_steps: 2
+    num_inference_timesteps: 4
+    add_pos_embed: true
+    max_seq_len: 1024
+    num_target_vision_tokens: 32
+    noise_beta_alpha: 1.5
+    noise_beta_beta: 1.0
+    noise_s: 0.999
+    num_timestep_buckets: 1000
+    diffusion_model_cfg:
+      action_dit_hidden_dim: 1024
+      dropout: 0.2
+      final_dropout: true
+      interleave_self_attention: true
+      norm_type: ada_norm
+      positional_embeddings: null
+      attention_head_dim: 64
+      output_dim: 1024
+      num_layers: 36
+      input_embedding_dim: 1024
+      cross_attention_dim: 1024
+      num_attention_heads: 16
+    loss_type: l1
+    future_action_window_size: 0
+    past_action_window_size: 0
+    action_env_dim: 2
+  kv_memory:
+    enabled: false
+    window: 4
+    rollout_len: 8
+    packed_train: false
+    rebased_sink: true
+version_id: '0.21'
+datasets:
+  vla_data:
+    include_state: true
+    obs_image_size: null
+    image_mode: single
+    num_obs_frames: 1
+    robot_type: rl_games_flappy
+trainer:
+  pretrained_checkpoint: null

zero-latency/flappy/vla/starvla-qwenpi_v3-h1/flappy-qwenpi_v3-hist8-zero-scratch-s42-v2/dataset_statistics.json

@@ -0,0 +1,194 @@
+{
+  "new_embodiment": {
+    "action": {
+      "mean": [
+        0.9167993068695068,
+        0.08320070058107376
+      ],
+      "std": [
+        0.2770693600177765,
+        0.2770693600177765
+      ],
+      "max": [
+        1.0,
+        1.0
+      ],
+      "min": [
+        0.0,
+        0.0
+      ],
+      "q01": [
+        0.0,
+        0.0
+      ],
+      "q99": [
+        1.0,
+        1.0
+      ],
+      "mask": [
+        true,
+        true
+      ]
+    },
+    "state": {
+      "mean": [
+        0.9977770447731018,
+        0.08308491855859756,
+        0.08308491855859756,
+        0.9980549216270447,
+        0.08311116695404053,
+        0.08311116695404053,
+        0.9983327984809875,
+        0.08314204216003418,
+        0.08314204216003418,
+        0.9986106753349304,
+        0.083157479763031,
+        0.083157479763031,
+        0.9988884925842285,
+        0.08316056430339813,
+        0.08316056430339813,
+        0.9991663694381714,
+        0.08316364884376526,
+        0.08316364884376526,
+        0.9994442462921143,
+        0.08316828310489655,
+        0.08316828310489655,
+        0.9997221231460571,
+        0.08318217843770981,
+        0.08318217843770981
+      ],
+      "std": [
+        0.04704940319061279,
+        0.2769029140472412,
+        0.2769029140472412,
+        0.04401940107345581,
+        0.2769410312175751,
+        0.2769410312175751,
+        0.04076549410820007,
+        0.2769874632358551,
+        0.2769874632358551,
+        0.03722505271434784,
+        0.27701058983802795,
+        0.27701058983802795,
+        0.033302940428256156,
+        0.2770152688026428,
+        0.2770152688026428,
+        0.028848618268966675,
+        0.2770198881626129,
+        0.2770198881626129,
+        0.023561282083393247,
+        0.27702683210372925,
+        0.27702683210372925,
+        0.016664866358041347,
+        0.27704158425331116,
+        0.27704158425331116
+      ],
+      "max": [
+        1.0,
+        1.0,
+        1.0,
+        1.0,
+        1.0,
+        1.0,
+        1.0,
+        1.0,
+        1.0,
+        1.0,
+        1.0,
+        1.0,
+        1.0,
+        1.0,
+        1.0,
+        1.0,
+        1.0,
+        1.0,
+        1.0,
+        1.0,
+        1.0,
+        1.0,
+        1.0,
+        1.0
+      ],
+      "min": [
+        0.0,
+        0.0,
+        0.0,
+        0.0,
+        0.0,
+        0.0,
+        0.0,
+        0.0,
+        0.0,
+        0.0,
+        0.0,
+        0.0,
+        0.0,
+        0.0,
+        0.0,
+        0.0,
+        0.0,
+        0.0,
+        0.0,
+        0.0,
+        0.0,
+        0.0,
+        0.0,
+        0.0
+      ],
+      "q01": [
+        1.0,
+        0.0,
+        0.0,
+        1.0,
+        0.0,
+        0.0,
+        1.0,
+        0.0,
+        0.0,
+        1.0,
+        0.0,
+        0.0,
+        1.0,
+        0.0,
+        0.0,
+        1.0,
+        0.0,
+        0.0,
+        1.0,
+        0.0,
+        0.0,
+        1.0,
+        0.0,
+        0.0
+      ],
+      "q99": [
+        1.0,
+        1.0,
+        1.0,
+        1.0,
+        1.0,
+        1.0,
+        1.0,
+        1.0,
+        1.0,
+        1.0,
+        1.0,
+        1.0,
+        1.0,
+        1.0,
+        1.0,
+        1.0,
+        1.0,
+        1.0,
+        1.0,
+        1.0,
+        1.0,
+        1.0,
+        1.0,
+        1.0
+      ]
+    },
+    "num_transitions": 647783,
+    "num_trajectories": 180
+  }
+}

zero-latency/flappy/vla/starvla-qwenpi_v3-h1/flappy-qwenpi_v3-hist8-zero-scratch-s42-v2/eval_l0.json

@@ -0,0 +1,42 @@
+{
+  "job_id": "eval_zero_to_zero__flappy-qwenpi_v3-hist8-zero-scratch-s42-v2",
+  "eval_protocol_version": "hist8-history01-par32-v3",
+  "training_step": 5000,
+  "checkpoint_sha256": "f6bf2adb10f48fba4c66564bf278966aad004231ef354b10ffb1a6b7e513fc60",
+  "profile_sha256": null,
+  "mean_return": 408.7170048055798,
+  "std_return": 88.4672673584004,
+  "mean_length": 3311.57,
+  "episodes": 100,
+  "parallel_envs": 32,
+  "max_episode_raw_frames": 3600,
+  "issued_actions": 331157,
+  "top_action": "noop",
+  "largest_action_fraction": 0.9169095021394685,
+  "unique_action_values": 2,
+  "episode_final_done": 19,
+  "episodes_at_frame_cap": 81,
+  "checks": {
+    "100_exact_seeds": true,
+    "checkpoint_sha": true,
+    "bundle_step": true,
+    "training_step_matches_plan": true,
+    "profile_sha": true,
+    "binary_history": true,
+    "raw_score_means_match": true,
+    "action_counts_match": true,
+    "action_total_match": true,
+    "summary_returns_match": true,
+    "summary_lengths_match": true,
+    "100_final_steps": true,
+    "final_step_lengths_match": true,
+    "parallel_capacity_horizon": true,
+    "fps_history": true,
+    "environment_latency_matches": true,
+    "episode_latency_protocol": true,
+    "training_recipe": true,
+    "training_dimensions": true,
+    "initialization_revision": true
+  },
+  "environment_condition": "zero"
+}

zero-latency/flappy/vla/starvla-qwenpi_v3-h1/flappy-qwenpi_v3-hist8-zero-scratch-s42-v2/eval_profile.json

@@ -0,0 +1,43 @@
+{
+  "job_id": "eval_zero_to_profile__flappy-qwenpi_v3-hist8-zero-scratch-s42-v2",
+  "eval_protocol_version": "hist8-history01-par32-v3",
+  "training_step": 5000,
+  "checkpoint_sha256": "f6bf2adb10f48fba4c66564bf278966aad004231ef354b10ffb1a6b7e513fc60",
+  "profile_sha256": "be2cd7a2a7fdf40c5c967d0c3cc16b912969d0ac0b4acc785aac3d0e4133df30",
+  "mean_return": 5.582000090926885,
+  "std_return": 1.9567003052385052,
+  "mean_length": 62.5,
+  "episodes": 100,
+  "parallel_envs": 32,
+  "max_episode_raw_frames": 3600,
+  "issued_actions": 2708,
+  "top_action": "noop",
+  "largest_action_fraction": 0.7828655834564254,
+  "unique_action_values": 2,
+  "episode_final_done": 100,
+  "episodes_at_frame_cap": 0,
+  "checks": {
+    "100_exact_seeds": true,
+    "checkpoint_sha": true,
+    "bundle_step": true,
+    "profile_sha": true,
+    "binary_history": true,
+    "raw_score_means_match": true,
+    "action_counts_match": true,
+    "action_total_match": true,
+    "summary_returns_match": true,
+    "summary_lengths_match": true,
+    "100_final_steps": true,
+    "final_step_lengths_match": true,
+    "training_condition_zero": true,
+    "environment_latency_matches": true,
+    "parallel_capacity_horizon": true,
+    "same_profile_as_profile_trained_eval": true,
+    "fps_history": true,
+    "episode_latency_protocol": true,
+    "training_recipe": true,
+    "training_dimensions": true,
+    "initialization_revision": true
+  },
+  "environment_condition": "profile"
+}

zero-latency/flappy/vla/starvla-qwenpi_v3-h1/flappy-qwenpi_v3-hist8-zero-scratch-s42-v2/evaluations/profile/action_audit.json

@@ -0,0 +1,125 @@
+{
+  "eval_protocol_version": "hist8-history01-par32-v3",
+  "history_encoding": "binary_0_1",
+  "episodes": 100,
+  "episode_seeds": [
+    1000000,
+    1000001,
+    1000002,
+    1000003,
+    1000004,
+    1000005,
+    1000006,
+    1000007,
+    1000008,
+    1000009,
+    1000010,
+    1000011,
+    1000012,
+    1000013,
+    1000014,
+    1000015,
+    1000016,
+    1000017,
+    1000018,
+    1000019,
+    1000020,
+    1000021,
+    1000022,
+    1000023,
+    1000024,
+    1000025,
+    1000026,
+    1000027,
+    1000028,
+    1000029,
+    1000030,
+    1000031,
+    1000032,
+    1000033,
+    1000034,
+    1000035,
+    1000036,
+    1000037,
+    1000038,
+    1000039,
+    1000040,
+    1000041,
+    1000042,
+    1000043,
+    1000044,
+    1000045,
+    1000046,
+    1000047,
+    1000048,
+    1000049,
+    1000050,
+    1000051,
+    1000052,
+    1000053,
+    1000054,
+    1000055,
+    1000056,
+    1000057,
+    1000058,
+    1000059,
+    1000060,
+    1000061,
+    1000062,
+    1000063,
+    1000064,
+    1000065,
+    1000066,
+    1000067,
+    1000068,
+    1000069,
+    1000070,
+    1000071,
+    1000072,
+    1000073,
+    1000074,
+    1000075,
+    1000076,
+    1000077,
+    1000078,
+    1000079,
+    1000080,
+    1000081,
+    1000082,
+    1000083,
+    1000084,
+    1000085,
+    1000086,
+    1000087,
+    1000088,
+    1000089,
+    1000090,
+    1000091,
+    1000092,
+    1000093,
+    1000094,
+    1000095,
+    1000096,
+    1000097,
+    1000098,
+    1000099
+  ],
+  "issued_actions": 2708,
+  "action_counts": {
+    "0": 2120,
+    "1": 588
+  },
+  "largest_action_fraction": 0.7828655834564254,
+  "raw_score_min": [
+    -0.08816230297088623,
+    -0.03477594256401062
+  ],
+  "raw_score_max": [
+    1.0449037551879883,
+    1.106244683265686
+  ],
+  "raw_score_mean": [
+    0.7854471598109409,
+    0.21968235313807075
+  ]
+}

zero-latency/flappy/vla/starvla-qwenpi_v3-h1/flappy-qwenpi_v3-hist8-zero-scratch-s42-v2/evaluations/profile/eval_config.yaml

@@ -0,0 +1,76 @@
+experiment:
+  name: flappy-qwenpi_v3-hist8-zero-scratch-s42-v2-step5000-history01-profile-par32
+  seed: 1000000
+backend:
+  run_mode: eval
+executor:
+  mode: simulated
+  simulated_worker_capacity: 1
+  simulated_inference_pool: true
+  inference_devices:
+  - cuda:0
+  - cuda:1
+  inference_batch_size: 16
+env:
+  name: flappy
+  gym_id: FlappyBird-v0
+  env_fps: 10
+  obs_fps: 10
+  render_mode: rgb_array
+  observation_mode: image
+  use_lidar: false
+  normalize_obs: true
+  audio_on: false
+  frame_stack: 1
+  obs_resize:
+  - 224
+  - 224
+  simulator: gpu
+  use_gpu_render: false
+  gpu_render_device: auto
+  gpu_render_profile: false
+  gpu_render_profile_interval: 200
+  noop_action: noop
+  action_map:
+    noop: 0
+    flap: 1
+  oneshot_actions:
+  - flap
+  action_history_decisions: 8
+latency:
+  method: temporal
+  profile_path: /home/ubuntu/codex/hist8-vla-20260923/results/profiling/profiles/pi05/1x-rtx3090/flappy/instance_a5037b165aa0cedc/profile.json
+  profile_worker_slot: 0
+  seed: 271828
+  add_latency_info: false
+scheduler:
+  hold_policy: one_frame_then_noop
+  ordering_policy: latest_ready
+policy:
+  type: starvla
+  checkpoint_path: /home/ubuntu/codex/hist8-vla-v2-20260923/outputs/tasks/decision_history_vla_hist8_v2/bundles/flappy-qwenpi_v3-hist8-zero-scratch-s42-v2/checkpoints/model.pt
+  model_config_path: /home/ubuntu/codex/hist8-vla-v2-20260923/outputs/tasks/decision_history_vla_hist8_v2/bundles/flappy-qwenpi_v3-hist8-zero-scratch-s42-v2/config.yaml
+  device: cuda:0
+  unnorm_key: new_embodiment
+  prompt_mode: latency_neutral
+  actions:
+  - noop
+  - flap
+  backbone_path: /home/ubuntu/.cache/huggingface/hub/models--Qwen--Qwen3-VL-4B-Instruct/snapshots/ebb281ec70b05090aa6165b016eac8ec08e71b17
+  worker_python_executable: /home/ubuntu/miniconda3/envs/starvla_rl_games_openvla/bin/python3.10
+  state_source: transport
+evaluation:
+  eval_episodes: 100
+  eval_parallel_envs: 32
+  latency_bench_env_backend: flappy_gpu_batched
+  eval_latency_values: null
+  eval_max_steps: 3600
+  eval_deterministic: true
+  eval_raw_reward: true
+logging:
+  output_dir: /home/ubuntu/codex/hist8-vla-v2-20260923/outputs/tasks/decision_history_vla_hist8_v2/evals/flappy-qwenpi_v3-hist8-zero-scratch-s42-v2/to_profile
+  save_step_records: true
+  save_action_records: true
+  save_latency_records: true
+  video:
+    enabled: false

zero-latency/flappy/vla/starvla-qwenpi_v3-h1/flappy-qwenpi_v3-hist8-zero-scratch-s42-v2/evaluations/profile/provenance.json

@@ -0,0 +1,46 @@
+{
+  "collection": {
+    "teacher_repo": "latency-sensitive-bench/benchmark-models",
+    "teacher_revision": "5b533937a72e9fa923ecf28b1c1be5a16bad488d",
+    "teacher_path": "zero-latency/flappy/small-policy/sample-factory-v1",
+    "teacher_checkpoint": "best_000047376_24281088_reward_2078.738.pth",
+    "teacher_sha256": "03e3ee5998a2bb661dcd396de6e0f29d9ffee2d1ab36c718af9f8c8c1c77734c",
+    "rollout_overrides": {
+      "env_fps": 10,
+      "obs_fps": 10,
+      "latency_type": "fixed",
+      "fixed_latency_ms": 0.0,
+      "seed": 42
+    },
+    "training_latency_condition": "zero",
+    "history": "decision_v1",
+    "history_decisions": 8,
+    "history_source": "zero_latency_issued_equals_applied",
+    "student_image_frames": 1,
+    "max_episode_raw_frames": 3600,
+    "episodes": 200,
+    "train_labels": 647783,
+    "filter_preset": "none"
+  },
+  "training_run_id": "flappy-qwenpi_v3-hist8-zero-scratch-s42-v2",
+  "training_step": 5000,
+  "global_batch": 128,
+  "base_model": "Qwen/Qwen3-VL-4B-Instruct",
+  "wandb_run_id": "9be1cc91",
+  "checkpoint_sha256": "f6bf2adb10f48fba4c66564bf278966aad004231ef354b10ffb1a6b7e513fc60",
+  "eval_protocol_version": "hist8-history01-par32-v3",
+  "history_encoding": "binary_0_1",
+  "episodes": 100,
+  "env_condition": "profile",
+  "parallel_envs": 32,
+  "max_episode_raw_frames": 3600,
+  "inference_devices": [
+    "cuda:0",
+    "cuda:1"
+  ],
+  "episode_seeds": [
+    1000000,
+    1000099
+  ],
+  "profile_sha256": "be2cd7a2a7fdf40c5c967d0c3cc16b912969d0ac0b4acc785aac3d0e4133df30"
+}

zero-latency/flappy/vla/starvla-qwenpi_v3-h1/flappy-qwenpi_v3-hist8-zero-scratch-s42-v2/evaluations/profile/resolved_config.yaml

@@ -0,0 +1,79 @@
+experiment:
+  name: flappy-qwenpi_v3-hist8-zero-scratch-s42-v2-step5000-history01-profile-par32
+  seed: 1000000
+backend:
+  run_mode: eval
+executor:
+  mode: simulated
+  simulated_worker_capacity: 1
+  simulated_inference_pool: true
+  inference_devices:
+  - cuda:0
+  - cuda:1
+  inference_batch_size: 16
+env:
+  name: flappy
+  gym_id: FlappyBird-v0
+  env_fps: 10
+  obs_fps: 10
+  render_mode: rgb_array
+  observation_mode: image
+  use_lidar: false
+  normalize_obs: true
+  audio_on: false
+  frame_stack: 1
+  obs_resize:
+  - 224
+  - 224
+  simulator: gpu
+  use_gpu_render: false
+  gpu_render_device: auto
+  gpu_render_profile: false
+  gpu_render_profile_interval: 200
+  noop_action: noop
+  action_map:
+    noop: 0
+    flap: 1
+  oneshot_actions:
+  - flap
+  action_history_decisions: 8
+latency:
+  method: temporal
+  profile_path: /home/ubuntu/codex/hist8-vla-20260923/results/profiling/profiles/pi05/1x-rtx3090/flappy/instance_a5037b165aa0cedc/profile.json
+  profile_worker_slot: 0
+  seed: 271828
+  add_latency_info: false
+scheduler:
+  hold_policy: one_frame_then_noop
+  ordering_policy: latest_ready
+policy:
+  type: starvla
+  checkpoint_path: /home/ubuntu/codex/hist8-vla-v2-20260923/outputs/tasks/decision_history_vla_hist8_v2/bundles/flappy-qwenpi_v3-hist8-zero-scratch-s42-v2/checkpoints/model.pt
+  model_config_path: /home/ubuntu/codex/hist8-vla-v2-20260923/outputs/tasks/decision_history_vla_hist8_v2/bundles/flappy-qwenpi_v3-hist8-zero-scratch-s42-v2/config.yaml
+  device: cuda:0
+  unnorm_key: new_embodiment
+  prompt_mode: latency_neutral
+  actions:
+  - noop
+  - flap
+  backbone_path: /home/ubuntu/.cache/huggingface/hub/models--Qwen--Qwen3-VL-4B-Instruct/snapshots/ebb281ec70b05090aa6165b016eac8ec08e71b17
+  worker_python_executable: /home/ubuntu/miniconda3/envs/starvla_rl_games_openvla/bin/python3.10
+  state_source: transport
+  action_prefix:
+    mode: none
+evaluation:
+  eval_episodes: 100
+  eval_parallel_envs: 32
+  latency_bench_env_backend: flappy_gpu_batched
+  eval_latency_values: null
+  eval_max_steps: 3600
+  eval_deterministic: true
+  eval_raw_reward: true
+logging:
+  output_dir: /home/ubuntu/codex/hist8-vla-v2-20260923/outputs/tasks/decision_history_vla_hist8_v2/evals/flappy-qwenpi_v3-hist8-zero-scratch-s42-v2/to_profile
+  save_step_records: true
+  save_action_records: true
+  save_latency_records: true
+  video:
+    enabled: false
+  simulated_pipeline_profile: false

zero-latency/flappy/vla/starvla-qwenpi_v3-h1/flappy-qwenpi_v3-hist8-zero-scratch-s42-v2/evaluations/profile/summary.json

@@ -0,0 +1,217 @@
+{
+  "checkpoint_path": "/home/ubuntu/codex/hist8-vla-v2-20260923/outputs/tasks/decision_history_vla_hist8_v2/bundles/flappy-qwenpi_v3-hist8-zero-scratch-s42-v2/checkpoints/model.pt",
+  "experiment_name": "flappy-qwenpi_v3-hist8-zero-scratch-s42-v2-step5000-history01-profile-par32",
+  "latency": "profile_sample",
+  "latency_type": "profile_sample",
+  "lengths": [
+    50,
+    70,
+    56,
+    63,
+    54,
+    122,
+    50,
+    50,
+    57,
+    93,
+    55,
+    50,
+    50,
+    50,
+    65,
+    50,
+    66,
+    50,
+    66,
+    70,
+    106,
+    55,
+    50,
+    50,
+    92,
+    50,
+    56,
+    50,
+    68,
+    58,
+    50,
+    65,
+    56,
+    50,
+    70,
+    86,
+    64,
+    67,
+    65,
+    62,
+    50,
+    64,
+    51,
+    86,
+    87,
+    92,
+    50,
+    50,
+    50,
+    89,
+    52,
+    53,
+    69,
+    57,
+    59,
+    70,
+    50,
+    64,
+    86,
+    102,
+    92,
+    68,
+    58,
+    41,
+    56,
+    54,
+    63,
+    52,
+    60,
+    69,
+    58,
+    68,
+    68,
+    67,
+    70,
+    50,
+    50,
+    50,
+    68,
+    57,
+    50,
+    57,
+    63,
+    50,
+    94,
+    86,
+    57,
+    50,
+    50,
+    65,
+    70,
+    31,
+    58,
+    58,
+    67,
+    69,
+    50,
+    68,
+    50,
+    50
+  ],
+  "mean_length": 62.5,
+  "mean_return": 5.582000090926885,
+  "returns": [
+    3.90000007301569,
+    6.800000101327896,
+    4.500000081956387,
+    6.100000090897083,
+    4.300000078976154,
+    12.900000177323818,
+    3.90000007301569,
+    3.90000007301569,
+    4.600000083446503,
+    9.100000135600567,
+    4.4000000804662704,
+    3.90000007301569,
+    3.90000007301569,
+    3.90000007301569,
+    6.3000000938773155,
+    3.90000007301569,
+    6.400000095367432,
+    3.90000007301569,
+    6.400000095367432,
+    6.800000101327896,
+    11.30000015348196,
+    4.4000000804662704,
+    3.90000007301569,
+    3.90000007301569,
+    9.00000013411045,
+    3.90000007301569,
+    4.500000081956387,
+    3.90000007301569,
+    6.600000098347664,
+    4.700000084936619,
+    3.90000007301569,
+    6.3000000938773155,
+    4.500000081956387,
+    3.90000007301569,
+    6.800000101327896,
+    8.400000125169754,
+    6.200000092387199,
+    6.500000096857548,
+    6.3000000938773155,
+    6.000000089406967,
+    3.90000007301569,
+    6.200000092387199,
+    4.000000074505806,
+    8.400000125169754,
+    8.50000012665987,
+    9.00000013411045,
+    3.90000007301569,
+    3.90000007301569,
+    3.90000007301569,
+    8.700000129640102,
+    4.100000075995922,
+    4.200000077486038,
+    6.70000009983778,
+    4.600000083446503,
+    4.800000086426735,
+    6.800000101327896,
+    3.90000007301569,
+    6.200000092387199,
+    8.400000125169754,
+    10.900000147521496,
+    9.00000013411045,
+    6.600000098347664,
+    4.700000084936619,
+    3.0000000596046448,
+    4.500000081956387,
+    4.300000078976154,
+    6.100000090897083,
+    4.100000075995922,
+    4.900000087916851,
+    6.70000009983778,
+    4.700000084936619,
+    6.600000098347664,
+    6.600000098347664,
+    6.500000096857548,
+    6.800000101327896,
+    3.90000007301569,
+    3.90000007301569,
+    3.90000007301569,
+    6.600000098347664,
+    4.600000083446503,
+    3.90000007301569,
+    4.600000083446503,
+    6.100000090897083,
+    3.90000007301569,
+    9.200000137090683,
+    8.400000125169754,
+    4.600000083446503,
+    3.90000007301569,
+    3.90000007301569,
+    6.3000000938773155,
+    6.800000101327896,
+    2.0000000447034836,
+    4.700000084936619,
+    4.700000084936619,
+    6.500000096857548,
+    6.70000009983778,
+    3.90000007301569,
+    6.600000098347664,
+    3.90000007301569,
+    3.90000007301569
+  ],
+  "seed": 1000000,
+  "source_profile_path": "/home/ubuntu/codex/hist8-vla-20260923/results/profiling/profiles/pi05/1x-rtx3090/flappy/instance_a5037b165aa0cedc/profile.json",
+  "std_return": 1.956700305238506,
+  "suite_name": "profile_sample",
+  "timestamp_utc": "2026-09-23T22:26:54.790529+00:00"
+}

zero-latency/flappy/vla/starvla-qwenpi_v3-h1/flappy-qwenpi_v3-hist8-zero-scratch-s42-v2/evaluations/profile/verification.json

@@ -0,0 +1,68 @@
+{
+  "job_id": "eval_zero_to_profile__flappy-qwenpi_v3-hist8-zero-scratch-s42-v2",
+  "task": "flappy",
+  "model": "qwenpi_v3",
+  "remote_eval_dir": "/home/ubuntu/codex/hist8-vla-v2-20260923/outputs/tasks/decision_history_vla_hist8_v2/evals/flappy-qwenpi_v3-hist8-zero-scratch-s42-v2/to_profile",
+  "checked_at": 1790202625.8180003,
+  "training_step": 5000,
+  "checkpoint_sha256": "f6bf2adb10f48fba4c66564bf278966aad004231ef354b10ffb1a6b7e513fc60",
+  "profile_sha256": "be2cd7a2a7fdf40c5c967d0c3cc16b912969d0ac0b4acc785aac3d0e4133df30",
+  "wandb_run_id": "9be1cc91",
+  "eval_protocol_version": "hist8-history01-par32-v3",
+  "max_episode_raw_frames": 3600,
+  "parallel_envs": 32,
+  "mean_return": 5.582000090926885,
+  "std_return": 1.9567003052385052,
+  "mean_length": 62.5,
+  "episodes": 100,
+  "issued_actions": 2708,
+  "top_action": "noop",
+  "largest_action_fraction": 0.7828655834564254,
+  "unique_action_values": 2,
+  "episode_final_done": 100,
+  "episodes_at_frame_cap": 0,
+  "checks": {
+    "100_exact_seeds": true,
+    "checkpoint_sha": true,
+    "bundle_step": true,
+    "profile_sha": true,
+    "binary_history": true,
+    "raw_score_means_match": true,
+    "action_counts_match": true,
+    "action_total_match": true,
+    "summary_returns_match": true,
+    "summary_lengths_match": true,
+    "100_final_steps": true,
+    "final_step_lengths_match": true,
+    "training_condition_zero": true,
+    "environment_latency_matches": true,
+    "parallel_capacity_horizon": true,
+    "same_profile_as_profile_trained_eval": true,
+    "fps_history": true,
+    "episode_latency_protocol": true,
+    "training_recipe": true,
+    "training_dimensions": true,
+    "initialization_revision": true
+  },
+  "job": {
+    "id": "eval_zero_to_profile__flappy-qwenpi_v3-hist8-zero-scratch-s42-v2",
+    "kind": "eval_ready",
+    "source": "v2",
+    "task": "flappy",
+    "model": "qwenpi_v3",
+    "python": "/home/ubuntu/miniconda3/envs/starvla_rl_games_openvla/bin/python",
+    "backbone": "/home/ubuntu/.cache/huggingface/hub/models--Qwen--Qwen3-VL-4B-Instruct/snapshots/ebb281ec70b05090aa6165b016eac8ec08e71b17",
+    "bundle_dir": "/home/ubuntu/codex/hist8-vla-v2-20260923/outputs/tasks/decision_history_vla_hist8_v2/bundles/flappy-qwenpi_v3-hist8-zero-scratch-s42-v2",
+    "profile_path": "/home/ubuntu/codex/hist8-vla-20260923/results/profiling/profiles/pi05/1x-rtx3090/flappy/instance_a5037b165aa0cedc/profile.json",
+    "env_condition": "profile",
+    "parallel_envs": 32,
+    "metric_prefix": "eval_hist8_v3/zero_to_profile",
+    "eval_dir": "/home/ubuntu/codex/hist8-vla-v2-20260923/outputs/tasks/decision_history_vla_hist8_v2/evals/flappy-qwenpi_v3-hist8-zero-scratch-s42-v2/to_profile",
+    "gpus_required": 2,
+    "priority": 1112,
+    "depends_on": [],
+    "ready_path": "/home/ubuntu/codex/hist8-vla-v2-20260923/outputs/tasks/decision_history_vla_hist8_v2/bundles/flappy-qwenpi_v3-hist8-zero-scratch-s42-v2/bundle.complete.json"
+  },
+  "evidence_dir": "/home/lixinyuan/latency-sensitive-bench/outputs/tasks/decision_history_vla_hist8_v2/evaluation_receipts/h200-2/eval_zero_to_profile__flappy-qwenpi_v3-hist8-zero-scratch-s42-v2",
+  "verification_status": "verified"
+}

zero-latency/flappy/vla/starvla-qwenpi_v3-h1/flappy-qwenpi_v3-hist8-zero-scratch-s42-v2/evaluations/zero/action_audit.json

@@ -0,0 +1,125 @@
+{
+  "eval_protocol_version": "hist8-history01-par32-v3",
+  "history_encoding": "binary_0_1",
+  "episodes": 100,
+  "episode_seeds": [
+    1000000,
+    1000001,
+    1000002,
+    1000003,
+    1000004,
+    1000005,
+    1000006,
+    1000007,
+    1000008,
+    1000009,
+    1000010,
+    1000011,
+    1000012,
+    1000013,
+    1000014,
+    1000015,
+    1000016,
+    1000017,
+    1000018,
+    1000019,
+    1000020,
+    1000021,
+    1000022,
+    1000023,
+    1000024,
+    1000025,
+    1000026,
+    1000027,
+    1000028,
+    1000029,
+    1000030,
+    1000031,
+    1000032,
+    1000033,
+    1000034,
+    1000035,
+    1000036,
+    1000037,
+    1000038,
+    1000039,
+    1000040,
+    1000041,
+    1000042,
+    1000043,
+    1000044,
+    1000045,
+    1000046,
+    1000047,
+    1000048,
+    1000049,
+    1000050,
+    1000051,
+    1000052,
+    1000053,
+    1000054,
+    1000055,
+    1000056,
+    1000057,
+    1000058,
+    1000059,
+    1000060,
+    1000061,
+    1000062,
+    1000063,
+    1000064,
+    1000065,
+    1000066,
+    1000067,
+    1000068,
+    1000069,
+    1000070,
+    1000071,
+    1000072,
+    1000073,
+    1000074,
+    1000075,
+    1000076,
+    1000077,
+    1000078,
+    1000079,
+    1000080,
+    1000081,
+    1000082,
+    1000083,
+    1000084,
+    1000085,
+    1000086,
+    1000087,
+    1000088,
+    1000089,
+    1000090,
+    1000091,
+    1000092,
+    1000093,
+    1000094,
+    1000095,
+    1000096,
+    1000097,
+    1000098,
+    1000099
+  ],
+  "issued_actions": 331157,
+  "action_counts": {
+    "0": 303641,
+    "1": 27516
+  },
+  "largest_action_fraction": 0.9169095021394685,
+  "raw_score_min": [
+    -0.25391101837158203,
+    -0.1234670877456665
+  ],
+  "raw_score_max": [
+    1.1082050800323486,
+    1.279944658279419
+  ],
+  "raw_score_mean": [
+    0.9261272017967728,
+    0.07972848236298555
+  ]
+}

zero-latency/flappy/vla/starvla-qwenpi_v3-h1/flappy-qwenpi_v3-hist8-zero-scratch-s42-v2/evaluations/zero/eval_config.yaml

@@ -0,0 +1,71 @@
+experiment:
+  name: flappy-qwenpi_v3-hist8-zero-scratch-s42-v2-step5000-history01-zero-par32
+  seed: 1000000
+backend:
+  run_mode: eval
+executor:
+  mode: simulated
+  simulated_worker_capacity: 1
+  simulated_inference_pool: true
+  inference_devices:
+  - cuda:0
+  inference_batch_size: 32
+env:
+  name: flappy
+  gym_id: FlappyBird-v0
+  env_fps: 10
+  obs_fps: 10
+  render_mode: rgb_array
+  observation_mode: image
+  use_lidar: false
+  normalize_obs: true
+  audio_on: false
+  frame_stack: 1
+  obs_resize:
+  - 224
+  - 224
+  simulator: gpu
+  use_gpu_render: false
+  gpu_render_device: auto
+  gpu_render_profile: false
+  gpu_render_profile_interval: 200
+  noop_action: noop
+  action_map:
+    noop: 0
+    flap: 1
+  oneshot_actions:
+  - flap
+  action_history_decisions: 8
+latency:
+  method: zero
+scheduler:
+  hold_policy: one_frame_then_noop
+  ordering_policy: latest_ready
+policy:
+  type: starvla
+  checkpoint_path: /home/ubuntu/codex/hist8-vla-v2-20260923/outputs/tasks/decision_history_vla_hist8_v2/bundles/flappy-qwenpi_v3-hist8-zero-scratch-s42-v2/checkpoints/model.pt
+  model_config_path: /home/ubuntu/codex/hist8-vla-v2-20260923/outputs/tasks/decision_history_vla_hist8_v2/bundles/flappy-qwenpi_v3-hist8-zero-scratch-s42-v2/config.yaml
+  device: cuda:0
+  unnorm_key: new_embodiment
+  prompt_mode: latency_neutral
+  actions:
+  - noop
+  - flap
+  backbone_path: /home/ubuntu/.cache/huggingface/hub/models--Qwen--Qwen3-VL-4B-Instruct/snapshots/ebb281ec70b05090aa6165b016eac8ec08e71b17
+  worker_python_executable: /home/ubuntu/miniconda3/envs/starvla_rl_games_openvla/bin/python3.10
+  state_source: transport
+evaluation:
+  eval_episodes: 100
+  eval_parallel_envs: 32
+  latency_bench_env_backend: flappy_gpu_batched
+  eval_latency_values: null
+  eval_max_steps: 3600
+  eval_deterministic: true
+  eval_raw_reward: true
+logging:
+  output_dir: /home/ubuntu/codex/hist8-vla-v2-20260923/outputs/tasks/decision_history_vla_hist8_v2/evals/flappy-qwenpi_v3-hist8-zero-scratch-s42-v2/to_zero
+  save_step_records: true
+  save_action_records: true
+  save_latency_records: true
+  video:
+    enabled: false

zero-latency/flappy/vla/starvla-qwenpi_v3-h1/flappy-qwenpi_v3-hist8-zero-scratch-s42-v2/evaluations/zero/provenance.json

@@ -0,0 +1,45 @@
+{
+  "collection": {
+    "teacher_repo": "latency-sensitive-bench/benchmark-models",
+    "teacher_revision": "5b533937a72e9fa923ecf28b1c1be5a16bad488d",
+    "teacher_path": "zero-latency/flappy/small-policy/sample-factory-v1",
+    "teacher_checkpoint": "best_000047376_24281088_reward_2078.738.pth",
+    "teacher_sha256": "03e3ee5998a2bb661dcd396de6e0f29d9ffee2d1ab36c718af9f8c8c1c77734c",
+    "rollout_overrides": {
+      "env_fps": 10,
+      "obs_fps": 10,
+      "latency_type": "fixed",
+      "fixed_latency_ms": 0.0,
+      "seed": 42
+    },
+    "training_latency_condition": "zero",
+    "history": "decision_v1",
+    "history_decisions": 8,
+    "history_source": "zero_latency_issued_equals_applied",
+    "student_image_frames": 1,
+    "max_episode_raw_frames": 3600,
+    "episodes": 200,
+    "train_labels": 647783,
+    "filter_preset": "none"
+  },
+  "training_run_id": "flappy-qwenpi_v3-hist8-zero-scratch-s42-v2",
+  "training_step": 5000,
+  "global_batch": 128,
+  "base_model": "Qwen/Qwen3-VL-4B-Instruct",
+  "wandb_run_id": "9be1cc91",
+  "checkpoint_sha256": "f6bf2adb10f48fba4c66564bf278966aad004231ef354b10ffb1a6b7e513fc60",
+  "eval_protocol_version": "hist8-history01-par32-v3",
+  "history_encoding": "binary_0_1",
+  "episodes": 100,
+  "env_condition": "zero",
+  "parallel_envs": 32,
+  "max_episode_raw_frames": 3600,
+  "inference_devices": [
+    "cuda:0"
+  ],
+  "episode_seeds": [
+    1000000,
+    1000099
+  ],
+  "profile_sha256": null
+}

zero-latency/flappy/vla/starvla-qwenpi_v3-h1/flappy-qwenpi_v3-hist8-zero-scratch-s42-v2/evaluations/zero/resolved_config.yaml

@@ -0,0 +1,74 @@
+experiment:
+  name: flappy-qwenpi_v3-hist8-zero-scratch-s42-v2-step5000-history01-zero-par32
+  seed: 1000000
+backend:
+  run_mode: eval
+executor:
+  mode: simulated
+  simulated_worker_capacity: 1
+  simulated_inference_pool: true
+  inference_devices:
+  - cuda:0
+  inference_batch_size: 32
+env:
+  name: flappy
+  gym_id: FlappyBird-v0
+  env_fps: 10
+  obs_fps: 10
+  render_mode: rgb_array
+  observation_mode: image
+  use_lidar: false
+  normalize_obs: true
+  audio_on: false
+  frame_stack: 1
+  obs_resize:
+  - 224
+  - 224
+  simulator: gpu
+  use_gpu_render: false
+  gpu_render_device: auto
+  gpu_render_profile: false
+  gpu_render_profile_interval: 200
+  noop_action: noop
+  action_map:
+    noop: 0
+    flap: 1
+  oneshot_actions:
+  - flap
+  action_history_decisions: 8
+latency:
+  method: zero
+scheduler:
+  hold_policy: one_frame_then_noop
+  ordering_policy: latest_ready
+policy:
+  type: starvla
+  checkpoint_path: /home/ubuntu/codex/hist8-vla-v2-20260923/outputs/tasks/decision_history_vla_hist8_v2/bundles/flappy-qwenpi_v3-hist8-zero-scratch-s42-v2/checkpoints/model.pt
+  model_config_path: /home/ubuntu/codex/hist8-vla-v2-20260923/outputs/tasks/decision_history_vla_hist8_v2/bundles/flappy-qwenpi_v3-hist8-zero-scratch-s42-v2/config.yaml
+  device: cuda:0
+  unnorm_key: new_embodiment
+  prompt_mode: latency_neutral
+  actions:
+  - noop
+  - flap
+  backbone_path: /home/ubuntu/.cache/huggingface/hub/models--Qwen--Qwen3-VL-4B-Instruct/snapshots/ebb281ec70b05090aa6165b016eac8ec08e71b17
+  worker_python_executable: /home/ubuntu/miniconda3/envs/starvla_rl_games_openvla/bin/python3.10
+  state_source: transport
+  action_prefix:
+    mode: none
+evaluation:
+  eval_episodes: 100
+  eval_parallel_envs: 32
+  latency_bench_env_backend: flappy_gpu_batched
+  eval_latency_values: null
+  eval_max_steps: 3600
+  eval_deterministic: true
+  eval_raw_reward: true
+logging:
+  output_dir: /home/ubuntu/codex/hist8-vla-v2-20260923/outputs/tasks/decision_history_vla_hist8_v2/evals/flappy-qwenpi_v3-hist8-zero-scratch-s42-v2/to_zero
+  save_step_records: true
+  save_action_records: true
+  save_latency_records: true
+  video:
+    enabled: false
+  simulated_pipeline_profile: false

zero-latency/flappy/vla/starvla-qwenpi_v3-h1/flappy-qwenpi_v3-hist8-zero-scratch-s42-v2/evaluations/zero/summary.json

@@ -0,0 +1,216 @@
+{
+  "checkpoint_path": "/home/ubuntu/codex/hist8-vla-v2-20260923/outputs/tasks/decision_history_vla_hist8_v2/bundles/flappy-qwenpi_v3-hist8-zero-scratch-s42-v2/checkpoints/model.pt",
+  "experiment_name": "flappy-qwenpi_v3-hist8-zero-scratch-s42-v2-step5000-history01-zero-par32",
+  "latency": 0,
+  "latency_type": "zero",
+  "lengths": [
+    3600,
+    2546,
+    1819,
+    802,
+    2918,
+    3600,
+    3600,
+    3600,
+    3600,
+    3600,
+    3600,
+    3600,
+    3600,
+    3600,
+    3600,
+    3600,
+    3600,
+    3600,
+    3600,
+    3600,
+    1592,
+    1867,
+    3600,
+    3600,
+    3600,
+    1334,
+    3600,
+    3600,
+    3600,
+    3600,
+    3600,
+    3600,
+    1799,
+    3600,
+    3600,
+    3600,
+    3600,
+    3600,
+    3600,
+    3600,
+    3600,
+    3600,
+    3600,
+    3600,
+    3600,
+    3600,
+    3600,
+    3600,
+    3600,
+    3600,
+    2269,
+    3600,
+    3600,
+    3600,
+    1340,
+    3600,
+    3600,
+    3600,
+    3600,
+    3600,
+    3399,
+    3600,
+    2495,
+    3600,
+    3600,
+    3600,
+    3600,
+    1123,
+    2382,
+    3600,
+    3600,
+    3600,
+    3600,
+    3600,
+    3600,
+    3600,
+    3512,
+    3600,
+    3496,
+    3600,
+    3600,
+    3600,
+    3600,
+    106,
+    2551,
+    3600,
+    2207,
+    3600,
+    3600,
+    3600,
+    3600,
+    3600,
+    3600,
+    3600,
+    3600,
+    3600,
+    3600,
+    3600,
+    3600,
+    3600
+  ],
+  "mean_length": 3311.57,
+  "mean_return": 408.7170048055798,
+  "returns": [
+    444.6000052243471,
+    312.90000369399786,
+    223.10000263899565,
+    97.10000116378069,
+    359.1000042334199,
+    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,
+    195.00000230967999,
+    228.8000027090311,
+    444.6000052243471,
+    444.6000052243471,
+    444.6000052243471,
+    162.90000193566084,
+    444.6000052243471,
+    444.6000052243471,
+    444.6000052243471,
+    444.6000052243471,
+    444.6000052243471,
+    444.6000052243471,
+    221.10000260919333,
+    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,
+    278.9000032916665,
+    444.6000052243471,
+    444.6000052243471,
+    444.6000052243471,
+    163.50000194460154,
+    444.6000052243471,
+    444.6000052243471,
+    444.6000052243471,
+    444.6000052243471,
+    444.6000052243471,
+    418.90000493079424,
+    444.6000052243471,
+    306.90000361949205,
+    444.6000052243471,
+    444.6000052243471,
+    444.6000052243471,
+    444.6000052243471,
+    137.30000162869692,
+    292.9000034555793,
+    444.6000052243471,
+    444.6000052243471,
+    444.6000052243471,
+    444.6000052243471,
+    444.6000052243471,
+    444.6000052243471,
+    444.6000052243471,
+    432.900005094707,
+    444.6000052243471,
+    431.30000507086515,
+    444.6000052243471,
+    444.6000052243471,
+    444.6000052243471,
+    444.6000052243471,
+    11.30000015348196,
+    314.3000036999583,
+    444.6000052243471,
+    270.90000320225954,
+    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
+  ],
+  "seed": 1000000,
+  "std_return": 88.46726735840035,
+  "suite_name": "fixed_0",
+  "timestamp_utc": "2026-09-23T23:37:36.976435+00:00"
+}

zero-latency/flappy/vla/starvla-qwenpi_v3-h1/flappy-qwenpi_v3-hist8-zero-scratch-s42-v2/evaluations/zero/verification.json

@@ -0,0 +1,69 @@
+{
+  "verification_status": "verified",
+  "job_id": "eval_zero_to_zero__flappy-qwenpi_v3-hist8-zero-scratch-s42-v2",
+  "task": "flappy",
+  "model": "qwenpi_v3",
+  "remote_eval_dir": "/home/ubuntu/codex/hist8-vla-v2-20260923/outputs/tasks/decision_history_vla_hist8_v2/evals/flappy-qwenpi_v3-hist8-zero-scratch-s42-v2/to_zero",
+  "checked_at": 1790208004.4863355,
+  "training_step": 5000,
+  "checkpoint_sha256": "f6bf2adb10f48fba4c66564bf278966aad004231ef354b10ffb1a6b7e513fc60",
+  "profile_sha256": null,
+  "wandb_run_id": "9be1cc91",
+  "eval_protocol_version": "hist8-history01-par32-v3",
+  "max_episode_raw_frames": 3600,
+  "parallel_envs": 32,
+  "mean_return": 408.7170048055798,
+  "std_return": 88.4672673584004,
+  "mean_length": 3311.57,
+  "episodes": 100,
+  "issued_actions": 331157,
+  "top_action": "noop",
+  "largest_action_fraction": 0.9169095021394685,
+  "unique_action_values": 2,
+  "episode_final_done": 19,
+  "episodes_at_frame_cap": 81,
+  "checks": {
+    "100_exact_seeds": true,
+    "checkpoint_sha": true,
+    "bundle_step": true,
+    "training_step_matches_plan": true,
+    "profile_sha": true,
+    "binary_history": true,
+    "raw_score_means_match": true,
+    "action_counts_match": true,
+    "action_total_match": true,
+    "summary_returns_match": true,
+    "summary_lengths_match": true,
+    "100_final_steps": true,
+    "final_step_lengths_match": true,
+    "parallel_capacity_horizon": true,
+    "fps_history": true,
+    "environment_latency_matches": true,
+    "episode_latency_protocol": true,
+    "training_recipe": true,
+    "training_dimensions": true,
+    "initialization_revision": true
+  },
+  "job": {
+    "id": "eval_zero_to_zero__flappy-qwenpi_v3-hist8-zero-scratch-s42-v2",
+    "kind": "eval_ready",
+    "source": "v2",
+    "task": "flappy",
+    "model": "qwenpi_v3",
+    "python": "/home/ubuntu/miniconda3/envs/starvla_rl_games_openvla/bin/python",
+    "backbone": "/home/ubuntu/.cache/huggingface/hub/models--Qwen--Qwen3-VL-4B-Instruct/snapshots/ebb281ec70b05090aa6165b016eac8ec08e71b17",
+    "bundle_dir": "/home/ubuntu/codex/hist8-vla-v2-20260923/outputs/tasks/decision_history_vla_hist8_v2/bundles/flappy-qwenpi_v3-hist8-zero-scratch-s42-v2",
+    "profile_path": "/home/ubuntu/codex/hist8-vla-20260923/results/profiling/profiles/pi05/1x-rtx3090/flappy/instance_a5037b165aa0cedc/profile.json",
+    "env_condition": "zero",
+    "parallel_envs": 32,
+    "metric_prefix": "eval_hist8_v3/zero_to_zero",
+    "eval_dir": "/home/ubuntu/codex/hist8-vla-v2-20260923/outputs/tasks/decision_history_vla_hist8_v2/evals/flappy-qwenpi_v3-hist8-zero-scratch-s42-v2/to_zero",
+    "gpus_required": 1,
+    "priority": 1112,
+    "depends_on": [
+      "package_zero__flappy-qwenpi_v3-hist8-zero-scratch-s42-v2"
+    ],
+    "aux_gpu": 7
+  },
+  "evidence_dir": "/home/lixinyuan/latency-sensitive-bench/outputs/tasks/decision_history_vla_hist8_v2/evaluation_receipts/h200-1/eval_zero_to_zero__flappy-qwenpi_v3-hist8-zero-scratch-s42-v2"
+}

zero-latency/flappy/vla/starvla-qwenpi_v3-h1/flappy-qwenpi_v3-hist8-zero-scratch-s42-v2/manifest.json

@@ -0,0 +1,111 @@
+{
+  "dataset_name": "lerobot",
+  "env_name": "flappy",
+  "episodes": 180,
+  "frames": 647783,
+  "task_prompts": [
+    "You are playing Flappy Bird. Pass through the pipe gaps and stay alive. Choose the action: NOOP, FLAP."
+  ],
+  "format": "starvla_lerobot_v2_decision_v1",
+  "source": "/home/ubuntu/codex/hist8-vla-v2-20260923/outputs/tasks/decision_history_vla_hist8_v2/data_zero/flappy/raw",
+  "robot_type": "rl_games_flappy",
+  "action_layout": "flappy_discrete",
+  "source_action_layout": "flappy_discrete",
+  "action_labels": [
+    "NOOP",
+    "FLAP"
+  ],
+  "action_dim": 2,
+  "active_action_dim": 2,
+  "state_source": "decision_v1",
+  "state_dim": 24,
+  "state_labels": [
+    "decision_history_0",
+    "decision_history_1",
+    "decision_history_2",
+    "decision_history_3",
+    "decision_history_4",
+    "decision_history_5",
+    "decision_history_6",
+    "decision_history_7",
+    "decision_history_8",
+    "decision_history_9",
+    "decision_history_10",
+    "decision_history_11",
+    "decision_history_12",
+    "decision_history_13",
+    "decision_history_14",
+    "decision_history_15",
+    "decision_history_16",
+    "decision_history_17",
+    "decision_history_18",
+    "decision_history_19",
+    "decision_history_20",
+    "decision_history_21",
+    "decision_history_22",
+    "decision_history_23"
+  ],
+  "state_normalization": {
+    "min": [
+      0.0,
+      0.0,
+      0.0,
+      0.0,
+      0.0,
+      0.0,
+      0.0,
+      0.0,
+      0.0,
+      0.0,
+      0.0,
+      0.0,
+      0.0,
+      0.0,
+      0.0,
+      0.0,
+      0.0,
+      0.0,
+      0.0,
+      0.0,
+      0.0,
+      0.0,
+      0.0,
+      0.0
+    ],
+    "max": [
+      1.0,
+      1.0,
+      1.0,
+      1.0,
+      1.0,
+      1.0,
+      1.0,
+      1.0,
+      1.0,
+      1.0,
+      1.0,
+      1.0,
+      1.0,
+      1.0,
+      1.0,
+      1.0,
+      1.0,
+      1.0,
+      1.0,
+      1.0,
+      1.0,
+      1.0,
+      1.0,
+      1.0
+    ]
+  },
+  "action_history_decisions": 8,
+  "rows_unit": "decision_step",
+  "obs_fps": 10.0,
+  "obs_stride_raw_frames": 1,
+  "latency_prompt_map_path": "/home/ubuntu/codex/hist8-vla-v2-20260923/outputs/tasks/decision_history_vla_hist8_v2/data_zero/flappy/lerobot/latency_prompt_map.json",
+  "custom_mixtures_path": "/home/ubuntu/codex/hist8-vla-v2-20260923/outputs/tasks/decision_history_vla_hist8_v2/data_zero/flappy/_generated_mixtures/lerobot.json",
+  "validation_dataset_name": "lerobot__val",
+  "validation_episodes": 20,
+  "validation_frames": 70375
+}

zero-latency/flappy/vla/starvla-qwenpi_v3-h1/flappy-qwenpi_v3-hist8-zero-scratch-s42-v2/profiles/hardware.json

@@ -0,0 +1,40 @@
+{
+  "driver_version": "595.58.03",
+  "gpu_class": "1x-rtx3090",
+  "gpus": [
+    {
+      "name": "NVIDIA GeForce RTX 3090",
+      "slot": 0
+    }
+  ],
+  "instance_id": "instance_a5037b165aa0cedc",
+  "topology_links": [],
+  "torch": {
+    "backends": {
+      "cuda_cudnn_sdp_enabled": true,
+      "cuda_flash_sdp_enabled": true,
+      "cuda_math_sdp_enabled": true,
+      "cuda_matmul_allow_tf32": false,
+      "cuda_mem_efficient_sdp_enabled": true,
+      "cudnn_allow_tf32": true,
+      "cudnn_benchmark": false
+    },
+    "cuda_available": true,
+    "cuda_device_count": 1,
+    "cuda_version": "12.8",
+    "current_device": 0,
+    "current_device_name": "NVIDIA GeForce RTX 3090",
+    "device_properties": [
+      {
+        "index": 0,
+        "major": 8,
+        "minor": 6,
+        "multi_processor_count": 82,
+        "name": "NVIDIA GeForce RTX 3090",
+        "total_memory": 25298141184
+      }
+    ],
+    "float32_matmul_precision": "highest",
+    "version": "2.11.0+cu128"
+  }
+}

zero-latency/flappy/vla/starvla-qwenpi_v3-h1/flappy-qwenpi_v3-hist8-zero-scratch-s42-v2/profiles/latency_burst_model.json

@@ -0,0 +1,30 @@
+{
+  "burst_dwell_distribution": null,
+  "burst_dwell_lengths": [],
+  "burst_merge_gap_records": 30,
+  "burst_rank_processes": [],
+  "burst_slot_rank_templates": [],
+  "model_type": "hidden_regime",
+  "pre_worker_time_ms": null,
+  "regime_step_counts": {
+    "burst": 0,
+    "calm": 397
+  },
+  "regime_transition_counts": {
+    "burst": {
+      "burst": 0,
+      "calm": 0
+    },
+    "calm": {
+      "burst": 0,
+      "calm": 392
+    }
+  },
+  "reset_scope": "session",
+  "schema_version": 12,
+  "spike_median_multiplier": 1.25,
+  "spike_threshold_ms_by_worker_slot": {
+    "0": 236.15164577960968
+  },
+  "worker_count": 1
+}

zero-latency/flappy/vla/starvla-qwenpi_v3-h1/flappy-qwenpi_v3-hist8-zero-scratch-s42-v2/profiles/latency_distribution.json

@@ -0,0 +1,1027 @@
+{
+  "schema_version": 3,
+  "worker_slots": {
+    "0": {
+      "all": {
+        "count": 397,
+        "observation_to_action_latency_ms": {
+          "distribution_type": "inverse_cdf",
+          "latency_ms": [
+            187.75602054595947,
+            187.75602054595947,
+            187.75602054595947,
+            187.75602054595947,
+            187.83226014137267,
+            187.93520952701567,
+            188.03688999176026,
+            188.13411481380462,
+            188.231339635849,
+            188.26682265424728,
+            188.27619265174866,
+            188.30949172115325,
+            188.44899653434754,
+            189.40945741653442,
+            189.69003904819488,
+            190.25492133140563,
+            191.0099636554718,
+            191.37371057510376,
+            191.63439695835115,
+            191.82081809997558,
+            191.9861323070526,
+            192.13247919082642,
+            192.4810184907913,
+            192.63652135849,
+            192.8018400478363,
+            192.9218871116638,
+            192.9732520341873,
+            193.10236066818237,
+            193.21885433197022,
+            193.36163339614868,
+            193.5950873374939,
+            193.66431140899658,
+            193.81515934944153,
+            193.91827468872071,
+            194.0794870662689,
+            194.30969717025758,
+            194.44131791591644,
+            194.6471332550049,
+            194.75994489192962,
+            194.9565594291687,
+            195.12569007873535,
+            195.3350133895874,
+            195.37710391521455,
+            195.5579750442505,
+            195.70858334064485,
+            195.86865257263185,
+            196.00170044898988,
+            196.1385356235504,
+            196.30027056217193,
+            196.45638925552367,
+            196.60483879566192,
+            196.8031542301178,
+            197.04250413894653,
+            197.21880411148072,
+            197.34508431911468,
+            197.56447951316832,
+            197.742094540596,
+            198.04206745147704,
+            198.52315817832948,
+            198.71906134605408,
+            198.80382729053497,
+            199.30283498764038,
+            199.36643128871918,
+            199.63361293792724,
+            199.7740250682831,
+            199.81841430664062,
+            200.0126831293106,
+            200.27857627868653,
+            200.6171898317337,
+            200.88457544326783,
+            201.0481881093979,
+            201.1880308151245,
+            201.47494828224183,
+            201.81525077819825,
+            202.00161520957946,
+            202.06214977264403,
+            202.3225512266159,
+            202.8287430381775,
+            203.21788419246673,
+            203.57608374595642,
+            203.97169956684112,
+            204.50646963119507,
+            205.11433240890503,
+            205.31188141822815,
+            206.07202221393584,
+            206.25563491821288,
+            206.4892382621765,
+            206.73607429504395,
+            207.19506292819977,
+            207.5799489593506,
+            207.70686626434326,
+            207.92139105796815,
+            208.29619679927825,
+            208.78365730285643,
+            209.20089116573334,
+            210.0040128326416,
+            210.2935305118561,
+            210.8714827156067,
+            211.13516402244568,
+            211.82103313446044,
+            211.96355032444,
+            212.22488346099854,
+            212.36337050437928,
+            212.9155470085144,
+            213.01242901325227,
+            213.30201068878173,
+            213.8809885263443,
+            214.8879083442688,
+            215.6156429386139,
+            216.05795721054076,
+            217.32847361564637,
+            217.34094309329987,
+            217.4501692314148,
+            217.58119548130034,
+            217.81515714359284,
+            218.29249999284744,
+            218.76984284210204,
+            220.6335994367601,
+            222.892192308426,
+            224.56480264663696,
+            224.56480264663696,
+            224.56480264663696,
+            224.56480264663696
+          ],
+          "quantile_levels": [
+            0.0,
+            0.0001,
+            0.0005,
+            0.001,
+            0.002,
+            0.003,
+            0.004,
+            0.005,
+            0.006,
+            0.007,
+            0.008,
+            0.009,
+            0.01,
+            0.02,
+            0.03,
+            0.04,
+            0.05,
+            0.06,
+            0.07,
+            0.08,
+            0.09,
+            0.1,
+            0.11,
+            0.12,
+            0.13,
+            0.14,
+            0.15,
+            0.16,
+            0.17,
+            0.18,
+            0.19,
+            0.2,
+            0.21,
+            0.22,
+            0.23,
+            0.24,
+            0.25,
+            0.26,
+            0.27,
+            0.28,
+            0.29,
+            0.3,
+            0.31,
+            0.32,
+            0.33,
+            0.34,
+            0.35,
+            0.36,
+            0.37,
+            0.38,
+            0.39,
+            0.4,
+            0.41,
+            0.42,
+            0.43,
+            0.44,
+            0.45,
+            0.46,
+            0.47,
+            0.48,
+            0.49,
+            0.5,
+            0.51,
+            0.52,
+            0.53,
+            0.54,
+            0.55,
+            0.56,
+            0.57,
+            0.58,
+            0.59,
+            0.6,
+            0.61,
+            0.62,
+            0.63,
+            0.64,
+            0.65,
+            0.66,
+            0.67,
+            0.68,
+            0.69,
+            0.7,
+            0.71,
+            0.72,
+            0.73,
+            0.74,
+            0.75,
+            0.76,
+            0.77,
+            0.78,
+            0.79,
+            0.8,
+            0.81,
+            0.82,
+            0.83,
+            0.84,
+            0.85,
+            0.86,
+            0.87,
+            0.88,
+            0.89,
+            0.9,
+            0.91,
+            0.92,
+            0.93,
+            0.94,
+            0.95,
+            0.96,
+            0.97,
+            0.98,
+            0.99,
+            0.991,
+            0.992,
+            0.993,
+            0.994,
+            0.995,
+            0.996,
+            0.997,
+            0.998,
+            0.999,
+            0.9995,
+            0.9999,
+            1.0
+          ]
+        },
+        "spearman_rho": 0.9902790419542948,
+        "worker_service_time_ms": {
+          "distribution_type": "inverse_cdf",
+          "latency_ms": [
+            184.851975440979,
+            184.851975440979,
+            184.851975440979,
+            184.851975440979,
+            184.93146674633027,
+            185.0388070464134,
+            185.14413385772704,
+            185.24239057779312,
+            185.3406472978592,
+            185.4222120604515,
+            185.4967171421051,
+            185.55999927854538,
+            185.5734699869156,
+            186.56576988220215,
+            186.94663113594055,
+            187.33117692947388,
+            187.8231789112091,
+            188.30710098266601,
+            188.65142943382264,
+            188.93863531112672,
+            189.03877035617828,
+            189.10015487670898,
+            189.17648320674897,
+            189.40980823516847,
+            189.83681179523467,
+            190.03180967330934,
+            190.12901203632356,
+            190.26392262458802,
+            190.43206815242766,
+            190.5500379562378,
+            190.71069761753083,
+            190.87697758674622,
+            190.96741201877595,
+            191.07456970214844,
+            191.1704991054535,
+            191.3348473072052,
+            191.45189881324768,
+            191.69053689956664,
+            191.8700308418274,
+            192.035788230896,
+            192.08613749027253,
+            192.205251121521,
+            192.39817912578582,
+            192.5210274219513,
+            192.7438492155075,
+            193.03918478012085,
+            193.11329154968263,
+            193.29157629013062,
+            193.35646272182464,
+            193.55194200515749,
+            193.88924970626832,
+            194.10706882476808,
+            194.23809970378875,
+            194.42386789321898,
+            194.73488067626954,
+            194.8295351600647,
+            194.9371924877167,
+            195.12608194351196,
+            195.51668857097627,
+            195.65883117675781,
+            196.06324529647827,
+            196.22055006027222,
+            196.4921687555313,
+            196.76865724563598,
+            196.96945638656615,
+            197.12704219818116,
+            197.323499250412,
+            197.46007429122926,
+            197.8862152147293,
+            198.15339071273803,
+            198.44281109809876,
+            198.49902439117432,
+            198.82631464958192,
+            199.01242155075073,
+            199.14573033332826,
+            199.23965560913086,
+            199.50943048000335,
+            199.9217773628235,
+            200.1374461555481,
+            200.4188088130951,
+            200.82518770217897,
+            201.18839120864868,
+            201.51233775138854,
+            202.0418456840515,
+            202.683725771904,
+            203.12282960891724,
+            203.7856501340866,
+            204.1314724445343,
+            204.46249304771425,
+            204.7306516265869,
+            204.96170333862304,
+            205.15730891227722,
+            205.52443151474,
+            205.7483952140808,
+            205.90362362861634,
+            206.58012862205504,
+            207.17616543769836,
+            207.39526243209838,
+            208.23647666931151,
+            208.85318040847778,
+            209.24760127544403,
+            209.55709486007692,
+            209.9005552625656,
+            210.3381750869751,
+            210.79555583000183,
+            210.98525899887085,
+            211.13733308315278,
+            211.8839385318756,
+            212.99927528381346,
+            213.3959973526001,
+            214.2036056613922,
+            214.49116540813446,
+            214.60472968101502,
+            214.67909126996994,
+            214.84522978019714,
+            215.22836626529693,
+            215.6115027503967,
+            216.82290225219737,
+            218.27018248367315,
+            219.34197187423706,
+            219.34197187423706,
+            219.34197187423706,
+            219.34197187423706
+          ],
+          "quantile_levels": [
+            0.0,
+            0.0001,
+            0.0005,
+            0.001,
+            0.002,
+            0.003,
+            0.004,
+            0.005,
+            0.006,
+            0.007,
+            0.008,
+            0.009,
+            0.01,
+            0.02,
+            0.03,
+            0.04,
+            0.05,
+            0.06,
+            0.07,
+            0.08,
+            0.09,
+            0.1,
+            0.11,
+            0.12,
+            0.13,
+            0.14,
+            0.15,
+            0.16,
+            0.17,
+            0.18,
+            0.19,
+            0.2,
+            0.21,
+            0.22,
+            0.23,
+            0.24,
+            0.25,
+            0.26,
+            0.27,
+            0.28,
+            0.29,
+            0.3,
+            0.31,
+            0.32,
+            0.33,
+            0.34,
+            0.35,
+            0.36,
+            0.37,
+            0.38,
+            0.39,
+            0.4,
+            0.41,
+            0.42,
+            0.43,
+            0.44,
+            0.45,
+            0.46,
+            0.47,
+            0.48,
+            0.49,
+            0.5,
+            0.51,
+            0.52,
+            0.53,
+            0.54,
+            0.55,
+            0.56,
+            0.57,
+            0.58,
+            0.59,
+            0.6,
+            0.61,
+            0.62,
+            0.63,
+            0.64,
+            0.65,
+            0.66,
+            0.67,
+            0.68,
+            0.69,
+            0.7,
+            0.71,
+            0.72,
+            0.73,
+            0.74,
+            0.75,
+            0.76,
+            0.77,
+            0.78,
+            0.79,
+            0.8,
+            0.81,
+            0.82,
+            0.83,
+            0.84,
+            0.85,
+            0.86,
+            0.87,
+            0.88,
+            0.89,
+            0.9,
+            0.91,
+            0.92,
+            0.93,
+            0.94,
+            0.95,
+            0.96,
+            0.97,
+            0.98,
+            0.99,
+            0.991,
+            0.992,
+            0.993,
+            0.994,
+            0.995,
+            0.996,
+            0.997,
+            0.998,
+            0.999,
+            0.9995,
+            0.9999,
+            1.0
+          ]
+        }
+      },
+      "steady": {
+        "count": 397,
+        "observation_to_action_latency_ms": {
+          "distribution_type": "inverse_cdf",
+          "latency_ms": [
+            187.75602054595947,
+            187.75602054595947,
+            187.75602054595947,
+            187.75602054595947,
+            187.83226014137267,
+            187.93520952701567,
+            188.03688999176026,
+            188.13411481380462,
+            188.231339635849,
+            188.26682265424728,
+            188.27619265174866,
+            188.30949172115325,
+            188.44899653434754,
+            189.40945741653442,
+            189.69003904819488,
+            190.25492133140563,
+            191.0099636554718,
+            191.37371057510376,
+            191.63439695835115,
+            191.82081809997558,
+            191.9861323070526,
+            192.13247919082642,
+            192.4810184907913,
+            192.63652135849,
+            192.8018400478363,
+            192.9218871116638,
+            192.9732520341873,
+            193.10236066818237,
+            193.21885433197022,
+            193.36163339614868,
+            193.5950873374939,
+            193.66431140899658,
+            193.81515934944153,
+            193.91827468872071,
+            194.0794870662689,
+            194.30969717025758,
+            194.44131791591644,
+            194.6471332550049,
+            194.75994489192962,
+            194.9565594291687,
+            195.12569007873535,
+            195.3350133895874,
+            195.37710391521455,
+            195.5579750442505,
+            195.70858334064485,
+            195.86865257263185,
+            196.00170044898988,
+            196.1385356235504,
+            196.30027056217193,
+            196.45638925552367,
+            196.60483879566192,
+            196.8031542301178,
+            197.04250413894653,
+            197.21880411148072,
+            197.34508431911468,
+            197.56447951316832,
+            197.742094540596,
+            198.04206745147704,
+            198.52315817832948,
+            198.71906134605408,
+            198.80382729053497,
+            199.30283498764038,
+            199.36643128871918,
+            199.63361293792724,
+            199.7740250682831,
+            199.81841430664062,
+            200.0126831293106,
+            200.27857627868653,
+            200.6171898317337,
+            200.88457544326783,
+            201.0481881093979,
+            201.1880308151245,
+            201.47494828224183,
+            201.81525077819825,
+            202.00161520957946,
+            202.06214977264403,
+            202.3225512266159,
+            202.8287430381775,
+            203.21788419246673,
+            203.57608374595642,
+            203.97169956684112,
+            204.50646963119507,
+            205.11433240890503,
+            205.31188141822815,
+            206.07202221393584,
+            206.25563491821288,
+            206.4892382621765,
+            206.73607429504395,
+            207.19506292819977,
+            207.5799489593506,
+            207.70686626434326,
+            207.92139105796815,
+            208.29619679927825,
+            208.78365730285643,
+            209.20089116573334,
+            210.0040128326416,
+            210.2935305118561,
+            210.8714827156067,
+            211.13516402244568,
+            211.82103313446044,
+            211.96355032444,
+            212.22488346099854,
+            212.36337050437928,
+            212.9155470085144,
+            213.01242901325227,
+            213.30201068878173,
+            213.8809885263443,
+            214.8879083442688,
+            215.6156429386139,
+            216.05795721054076,
+            217.32847361564637,
+            217.34094309329987,
+            217.4501692314148,
+            217.58119548130034,
+            217.81515714359284,
+            218.29249999284744,
+            218.76984284210204,
+            220.6335994367601,
+            222.892192308426,
+            224.56480264663696,
+            224.56480264663696,
+            224.56480264663696,
+            224.56480264663696
+          ],
+          "quantile_levels": [
+            0.0,
+            0.0001,
+            0.0005,
+            0.001,
+            0.002,
+            0.003,
+            0.004,
+            0.005,
+            0.006,
+            0.007,
+            0.008,
+            0.009,
+            0.01,
+            0.02,
+            0.03,
+            0.04,
+            0.05,
+            0.06,
+            0.07,
+            0.08,
+            0.09,
+            0.1,
+            0.11,
+            0.12,
+            0.13,
+            0.14,
+            0.15,
+            0.16,
+            0.17,
+            0.18,
+            0.19,
+            0.2,
+            0.21,
+            0.22,
+            0.23,
+            0.24,
+            0.25,
+            0.26,
+            0.27,
+            0.28,
+            0.29,
+            0.3,
+            0.31,
+            0.32,
+            0.33,
+            0.34,
+            0.35,
+            0.36,
+            0.37,
+            0.38,
+            0.39,
+            0.4,
+            0.41,
+            0.42,
+            0.43,
+            0.44,
+            0.45,
+            0.46,
+            0.47,
+            0.48,
+            0.49,
+            0.5,
+            0.51,
+            0.52,
+            0.53,
+            0.54,
+            0.55,
+            0.56,
+            0.57,
+            0.58,
+            0.59,
+            0.6,
+            0.61,
+            0.62,
+            0.63,
+            0.64,
+            0.65,
+            0.66,
+            0.67,
+            0.68,
+            0.69,
+            0.7,
+            0.71,
+            0.72,
+            0.73,
+            0.74,
+            0.75,
+            0.76,
+            0.77,
+            0.78,
+            0.79,
+            0.8,
+            0.81,
+            0.82,
+            0.83,
+            0.84,
+            0.85,
+            0.86,
+            0.87,
+            0.88,
+            0.89,
+            0.9,
+            0.91,
+            0.92,
+            0.93,
+            0.94,
+            0.95,
+            0.96,
+            0.97,
+            0.98,
+            0.99,
+            0.991,
+            0.992,
+            0.993,
+            0.994,
+            0.995,
+            0.996,
+            0.997,
+            0.998,
+            0.999,
+            0.9995,
+            0.9999,
+            1.0
+          ]
+        },
+        "spearman_rho": 0.9902790419542948,
+        "worker_service_time_ms": {
+          "distribution_type": "inverse_cdf",
+          "latency_ms": [
+            184.851975440979,
+            184.851975440979,
+            184.851975440979,
+            184.851975440979,
+            184.93146674633027,
+            185.0388070464134,
+            185.14413385772704,
+            185.24239057779312,
+            185.3406472978592,
+            185.4222120604515,
+            185.4967171421051,
+            185.55999927854538,
+            185.5734699869156,
+            186.56576988220215,
+            186.94663113594055,
+            187.33117692947388,
+            187.8231789112091,
+            188.30710098266601,
+            188.65142943382264,
+            188.93863531112672,
+            189.03877035617828,
+            189.10015487670898,
+            189.17648320674897,
+            189.40980823516847,
+            189.83681179523467,
+            190.03180967330934,
+            190.12901203632356,
+            190.26392262458802,
+            190.43206815242766,
+            190.5500379562378,
+            190.71069761753083,
+            190.87697758674622,
+            190.96741201877595,
+            191.07456970214844,
+            191.1704991054535,
+            191.3348473072052,
+            191.45189881324768,
+            191.69053689956664,
+            191.8700308418274,
+            192.035788230896,
+            192.08613749027253,
+            192.205251121521,
+            192.39817912578582,
+            192.5210274219513,
+            192.7438492155075,
+            193.03918478012085,
+            193.11329154968263,
+            193.29157629013062,
+            193.35646272182464,
+            193.55194200515749,
+            193.88924970626832,
+            194.10706882476808,
+            194.23809970378875,
+            194.42386789321898,
+            194.73488067626954,
+            194.8295351600647,
+            194.9371924877167,
+            195.12608194351196,
+            195.51668857097627,
+            195.65883117675781,
+            196.06324529647827,
+            196.22055006027222,
+            196.4921687555313,
+            196.76865724563598,
+            196.96945638656615,
+            197.12704219818116,
+            197.323499250412,
+            197.46007429122926,
+            197.8862152147293,
+            198.15339071273803,
+            198.44281109809876,
+            198.49902439117432,
+            198.82631464958192,
+            199.01242155075073,
+            199.14573033332826,
+            199.23965560913086,
+            199.50943048000335,
+            199.9217773628235,
+            200.1374461555481,
+            200.4188088130951,
+            200.82518770217897,
+            201.18839120864868,
+            201.51233775138854,
+            202.0418456840515,
+            202.683725771904,
+            203.12282960891724,
+            203.7856501340866,
+            204.1314724445343,
+            204.46249304771425,
+            204.7306516265869,
+            204.96170333862304,
+            205.15730891227722,
+            205.52443151474,
+            205.7483952140808,
+            205.90362362861634,
+            206.58012862205504,
+            207.17616543769836,
+            207.39526243209838,
+            208.23647666931151,
+            208.85318040847778,
+            209.24760127544403,
+            209.55709486007692,
+            209.9005552625656,
+            210.3381750869751,
+            210.79555583000183,
+            210.98525899887085,
+            211.13733308315278,
+            211.8839385318756,
+            212.99927528381346,
+            213.3959973526001,
+            214.2036056613922,
+            214.49116540813446,
+            214.60472968101502,
+            214.67909126996994,
+            214.84522978019714,
+            215.22836626529693,
+            215.6115027503967,
+            216.82290225219737,
+            218.27018248367315,
+            219.34197187423706,
+            219.34197187423706,
+            219.34197187423706,
+            219.34197187423706
+          ],
+          "quantile_levels": [
+            0.0,
+            0.0001,
+            0.0005,
+            0.001,
+            0.002,
+            0.003,
+            0.004,
+            0.005,
+            0.006,
+            0.007,
+            0.008,
+            0.009,
+            0.01,
+            0.02,
+            0.03,
+            0.04,
+            0.05,
+            0.06,
+            0.07,
+            0.08,
+            0.09,
+            0.1,
+            0.11,
+            0.12,
+            0.13,
+            0.14,
+            0.15,
+            0.16,
+            0.17,
+            0.18,
+            0.19,
+            0.2,
+            0.21,
+            0.22,
+            0.23,
+            0.24,
+            0.25,
+            0.26,
+            0.27,
+            0.28,
+            0.29,
+            0.3,
+            0.31,
+            0.32,
+            0.33,
+            0.34,
+            0.35,
+            0.36,
+            0.37,
+            0.38,
+            0.39,
+            0.4,
+            0.41,
+            0.42,
+            0.43,
+            0.44,
+            0.45,
+            0.46,
+            0.47,
+            0.48,
+            0.49,
+            0.5,
+            0.51,
+            0.52,
+            0.53,
+            0.54,
+            0.55,
+            0.56,
+            0.57,
+            0.58,
+            0.59,
+            0.6,
+            0.61,
+            0.62,
+            0.63,
+            0.64,
+            0.65,
+            0.66,
+            0.67,
+            0.68,
+            0.69,
+            0.7,
+            0.71,
+            0.72,
+            0.73,
+            0.74,
+            0.75,
+            0.76,
+            0.77,
+            0.78,
+            0.79,
+            0.8,
+            0.81,
+            0.82,
+            0.83,
+            0.84,
+            0.85,
+            0.86,
+            0.87,
+            0.88,
+            0.89,
+            0.9,
+            0.91,
+            0.92,
+            0.93,
+            0.94,
+            0.95,
+            0.96,
+            0.97,
+            0.98,
+            0.99,
+            0.991,
+            0.992,
+            0.993,
+            0.994,
+            0.995,
+            0.996,
+            0.997,
+            0.998,
+            0.999,
+            0.9995,
+            0.9999,
+            1.0
+          ]
+        }
+      }
+    }
+  }
+}

zero-latency/flappy/vla/starvla-qwenpi_v3-h1/flappy-qwenpi_v3-hist8-zero-scratch-s42-v2/profiles/profile.json

@@ -0,0 +1,66 @@
+{
+  "burst_model_path": "latency_burst_model.json",
+  "distribution_path": "latency_distribution.json",
+  "env_fps": 30,
+  "frame_ms": 33.333333333333336,
+  "gpu_class": "1x-rtx3090",
+  "instance_id": "instance_a5037b165aa0cedc",
+  "latency_kind": "observation_to_action_latency",
+  "latency_method": "temporal",
+  "model_id": "pi05",
+  "n_admitted_observations": 397,
+  "n_capacity_drops": 2004,
+  "n_observation_attempts": 2401,
+  "per_slot_summary": {
+    "0": {
+      "admitted_count": 397,
+      "mean_observation_to_action_latency_ms": 200.60598407404248,
+      "mean_worker_service_time_ms": 197.7043983581985,
+      "p95_observation_to_action_latency_ms": 213.87126626968384,
+      "p95_worker_service_time_ms": 211.10870447158814,
+      "p99_worker_service_time_ms": 213.84868305206302
+    }
+  },
+  "provenance": {
+    "base_config": "configs/experiments/flappy/realtime/starvla_profile.yaml",
+    "checkpoint_kind": "best",
+    "model_artifact": {
+      "checkpoint": "checkpoints/steps_5000_state/model.safetensors",
+      "model_config": "config.full.yaml",
+      "path_in_repo": ".",
+      "repo_id": "latency-sensitive-bench/pi05_flappy_fix_latency_0_200ep",
+      "source": "local"
+    },
+    "session_ids": [
+      0,
+      1,
+      2,
+      3,
+      4
+    ]
+  },
+  "sample_model_type": "hidden_regime",
+  "source_run_id": "20260914T125121248545Z",
+  "summary": {
+    "frame_ms": 33.333333333333336,
+    "max_ms": 224.56480264663696,
+    "mean_effective_frames": 6.018179522221274,
+    "mean_ms": 200.60598407404248,
+    "min_ms": 187.75602054595947,
+    "n_samples": 397,
+    "p50_frames": 5.979085049629211,
+    "p50_ms": 199.30283498764038,
+    "p90_frames": 6.36656447982788,
+    "p90_ms": 212.21881599426268,
+    "p95_frames": 6.416137988090515,
+    "p95_ms": 213.87126626968384,
+    "p99_frames": 6.519392492294311,
+    "p99_ms": 217.31308307647706,
+    "prob_latency_gt_1_frame": 1.0,
+    "prob_latency_gt_2_frames": 1.0,
+    "prob_latency_gt_3_frames": 1.0,
+    "std_ms": 7.519660172748261
+  },
+  "visualization_path": "latency_profile.png",
+  "workload_id": "flappy"
+}

zero-latency/flappy/vla/starvla-qwenpi_v3-h1/flappy-qwenpi_v3-hist8-zero-scratch-s42-v2/provenance.json

@@ -0,0 +1,124 @@
+{
+  "collection": {
+    "teacher_repo": "latency-sensitive-bench/benchmark-models",
+    "teacher_revision": "5b533937a72e9fa923ecf28b1c1be5a16bad488d",
+    "teacher_path": "zero-latency/flappy/small-policy/sample-factory-v1",
+    "teacher_checkpoint": "best_000047376_24281088_reward_2078.738.pth",
+    "teacher_sha256": "03e3ee5998a2bb661dcd396de6e0f29d9ffee2d1ab36c718af9f8c8c1c77734c",
+    "rollout_overrides": {
+      "env_fps": 10,
+      "obs_fps": 10,
+      "latency_type": "fixed",
+      "fixed_latency_ms": 0.0,
+      "seed": 42
+    },
+    "training_latency_condition": "zero",
+    "history": "decision_v1",
+    "history_decisions": 8,
+    "history_source": "zero_latency_issued_equals_applied",
+    "student_image_frames": 1,
+    "max_episode_raw_frames": 3600,
+    "episodes": 200,
+    "train_labels": 647783,
+    "filter_preset": "none"
+  },
+  "training_run_id": "flappy-qwenpi_v3-hist8-zero-scratch-s42-v2",
+  "training_step": 5000,
+  "global_batch": 128,
+  "base_model": "Qwen/Qwen3-VL-4B-Instruct",
+  "wandb_run_id": "9be1cc91",
+  "checkpoint_sha256": "f6bf2adb10f48fba4c66564bf278966aad004231ef354b10ffb1a6b7e513fc60",
+  "eval_protocol_version": "hist8-history01-par32-v3",
+  "history_encoding": "binary_0_1",
+  "training_condition": "zero",
+  "base_model_revision": "ebb281ec70b05090aa6165b016eac8ec08e71b17",
+  "hf_repo_id": "latency-sensitive-bench/benchmark-models",
+  "hf_path": "zero-latency/flappy/vla/starvla-qwenpi_v3-h1/flappy-qwenpi_v3-hist8-zero-scratch-s42-v2",
+  "evaluations": [
+    {
+      "job_id": "eval_zero_to_profile__flappy-qwenpi_v3-hist8-zero-scratch-s42-v2",
+      "eval_protocol_version": "hist8-history01-par32-v3",
+      "training_step": 5000,
+      "checkpoint_sha256": "f6bf2adb10f48fba4c66564bf278966aad004231ef354b10ffb1a6b7e513fc60",
+      "profile_sha256": "be2cd7a2a7fdf40c5c967d0c3cc16b912969d0ac0b4acc785aac3d0e4133df30",
+      "mean_return": 5.582000090926885,
+      "std_return": 1.9567003052385052,
+      "mean_length": 62.5,
+      "episodes": 100,
+      "parallel_envs": 32,
+      "max_episode_raw_frames": 3600,
+      "issued_actions": 2708,
+      "top_action": "noop",
+      "largest_action_fraction": 0.7828655834564254,
+      "unique_action_values": 2,
+      "episode_final_done": 100,
+      "episodes_at_frame_cap": 0,
+      "checks": {
+        "100_exact_seeds": true,
+        "checkpoint_sha": true,
+        "bundle_step": true,
+        "profile_sha": true,
+        "binary_history": true,
+        "raw_score_means_match": true,
+        "action_counts_match": true,
+        "action_total_match": true,
+        "summary_returns_match": true,
+        "summary_lengths_match": true,
+        "100_final_steps": true,
+        "final_step_lengths_match": true,
+        "training_condition_zero": true,
+        "environment_latency_matches": true,
+        "parallel_capacity_horizon": true,
+        "same_profile_as_profile_trained_eval": true,
+        "fps_history": true,
+        "episode_latency_protocol": true,
+        "training_recipe": true,
+        "training_dimensions": true,
+        "initialization_revision": true
+      },
+      "environment_condition": "profile"
+    },
+    {
+      "job_id": "eval_zero_to_zero__flappy-qwenpi_v3-hist8-zero-scratch-s42-v2",
+      "eval_protocol_version": "hist8-history01-par32-v3",
+      "training_step": 5000,
+      "checkpoint_sha256": "f6bf2adb10f48fba4c66564bf278966aad004231ef354b10ffb1a6b7e513fc60",
+      "profile_sha256": null,
+      "mean_return": 408.7170048055798,
+      "std_return": 88.4672673584004,
+      "mean_length": 3311.57,
+      "episodes": 100,
+      "parallel_envs": 32,
+      "max_episode_raw_frames": 3600,
+      "issued_actions": 331157,
+      "top_action": "noop",
+      "largest_action_fraction": 0.9169095021394685,
+      "unique_action_values": 2,
+      "episode_final_done": 19,
+      "episodes_at_frame_cap": 81,
+      "checks": {
+        "100_exact_seeds": true,
+        "checkpoint_sha": true,
+        "bundle_step": true,
+        "training_step_matches_plan": true,
+        "profile_sha": true,
+        "binary_history": true,
+        "raw_score_means_match": true,
+        "action_counts_match": true,
+        "action_total_match": true,
+        "summary_returns_match": true,
+        "summary_lengths_match": true,
+        "100_final_steps": true,
+        "final_step_lengths_match": true,
+        "parallel_capacity_horizon": true,
+        "fps_history": true,
+        "environment_latency_matches": true,
+        "episode_latency_protocol": true,
+        "training_recipe": true,
+        "training_dimensions": true,
+        "initialization_revision": true
+      },
+      "environment_condition": "zero"
+    }
+  ]
+}

zero-latency/flappy/vla/starvla-qwenpi_v3-h1/flappy-qwenpi_v3-hist8-zero-scratch-s42-v2/rollout_metadata.json

@@ -0,0 +1,338 @@
+{
+  "action_history_decisions": 8,
+  "action_labels": [
+    "NOOP",
+    "FLAP"
+  ],
+  "action_layout": "flappy_discrete",
+  "base_prompt": "You are playing Flappy Bird. Pass through the pipe gaps and stay alive. Choose the action: NOOP, FLAP.",
+  "checkpoint_config": {
+    "actor_critic_share_weights": true,
+    "actor_worker_gpus": [
+      0
+    ],
+    "adam_beta1": 0.9,
+    "adam_beta2": 0.999,
+    "adam_eps": 1e-05,
+    "adaptive_stddev": true,
+    "add_latency_info": false,
+    "algo": "APPO",
+    "async_rl": true,
+    "audio_on": false,
+    "background": "day",
+    "batch_size": 4096,
+    "batched_sampling": true,
+    "benchmark": false,
+    "bird_color": "yellow",
+    "cli_args": {
+      "actor_worker_gpus": [
+        0
+      ],
+      "add_latency_info": false,
+      "algo": "APPO",
+      "async_rl": true,
+      "audio_on": false,
+      "background": "day",
+      "batch_size": 4096,
+      "batched_sampling": true,
+      "bird_color": "yellow",
+      "debug": false,
+      "debug_timelimit_diagnostics": false,
+      "device": "gpu",
+      "env": "latency_flappy",
+      "env_fps": 30.0,
+      "env_gpu_actions": true,
+      "env_gpu_observations": true,
+      "episode_metrics_path": "results/checkpoints/jitter_destructiveness/real_profile/teachers_latency0/flappy_teacher_fixed_l0_obs30_seed0/episode_metrics.jsonl",
+      "eval_deterministic": true,
+      "eval_episodes": 100,
+      "eval_max_steps": 7200,
+      "eval_parallel_envs": 50,
+      "experiment": "flappy_teacher_fixed_l0_obs30_seed0",
+      "experiment_summaries_interval": 1,
+      "fixed_latency_ms": 0.0,
+      "frame_stack": 4,
+      "gae_lambda": 0.95,
+      "gamma": 0.99,
+      "gpu_render_batch_size": 128,
+      "gpu_render_device": "auto",
+      "gpu_render_profile": false,
+      "gpu_render_profile_interval": 200,
+      "gym_id": "FlappyBird-v0",
+      "hold_policy": "one_frame_then_noop",
+      "keep_checkpoints": 5,
+      "latency_type": "fixed",
+      "learning_rate": 0.00025,
+      "max_grad_norm": 0.5,
+      "max_policy_lag": 400,
+      "mode": "train",
+      "normalize_input": true,
+      "normalize_obs": true,
+      "normalize_returns": true,
+      "num_batches_per_epoch": 8,
+      "num_envs_per_worker": 1,
+      "num_epochs": 8,
+      "num_policies": 1,
+      "num_workers": 2,
+      "obs_fps": 30.0,
+      "obs_resize": "84,84",
+      "ordering_policy": "latest_ready",
+      "pipe_color": "green",
+      "pipe_gap": 100,
+      "ppo_clip_ratio": 0.1,
+      "ppo_clip_value": 0.2,
+      "recurrence": 1,
+      "restart_behavior": "resume",
+      "rollout": 128,
+      "save_every_sec": 600,
+      "seed": 0,
+      "simulator": "gpu",
+      "stats_avg": 100,
+      "train_dir": "results/checkpoints/jitter_destructiveness/real_profile/teachers_latency0",
+      "train_for_env_steps": 25000000,
+      "use_gpu_render": true,
+      "use_lidar": false,
+      "use_rnn": false,
+      "value_loss_coeff": 0.5,
+      "wandb_group": "flappy-obs30-fixed_l0",
+      "wandb_job_type": "sample_factory",
+      "wandb_project": "latency-sensitive-bench",
+      "wandb_tags": [
+        "real_profile_latency0_teacher",
+        "flappy",
+        "teacher",
+        "fixed",
+        "fixed_l0",
+        "obs30",
+        "seed0"
+      ],
+      "with_wandb": true,
+      "worker_num_splits": 1
+    },
+    "command_line": "--mode train --algo APPO --env latency_flappy --experiment flappy_teacher_fixed_l0_obs30_seed0 --train_dir results/checkpoints/jitter_destructiveness/real_profile/teachers_latency0 --restart_behavior resume --device gpu --actor_worker_gpus 0 --env_gpu_observations True --env_gpu_actions True --gpu-render-batch-size 128 --seed 0 --episode_metrics_path results/checkpoints/jitter_destructiveness/real_profile/teachers_latency0/flappy_teacher_fixed_l0_obs30_seed0/episode_metrics.jsonl --train_for_env_steps 25000000 --num_workers 2 --num_envs_per_worker 1 --num_policies 1 --batch_size 4096 --rollout 128 --recurrence 1 --num_epochs 8 --num_batches_per_epoch 8 --worker_num_splits 1 --max_policy_lag 400 --learning_rate 0.00025 --gamma 0.99 --gae_lambda 0.95 --ppo_clip_ratio 0.1 --ppo_clip_value 0.2 --value_loss_coeff 0.5 --max_grad_norm 0.5 --save_every_sec 600 --keep_checkpoints 5 --stats_avg 100 --experiment_summaries_interval 1 --batched_sampling True --async_rl True --use_rnn False --normalize_returns True --normalize_input True --latency-type fixed --fixed-latency-ms 0.0 --add-latency-info False --eval-episodes 100 --eval-parallel-envs 50 --eval-max-steps 7200 --eval-deterministic True --with_wandb True --wandb_project latency-sensitive-bench --wandb_group flappy-obs30-fixed_l0 --wandb_job_type sample_factory --wandb_tags real_profile_latency0_teacher flappy teacher fixed fixed_l0 obs30 seed0 --gym_id FlappyBird-v0 --env-fps 30 --obs-fps 30.0 --use_lidar False --normalize_obs True --audio_on False --obs_resize 84,84 --use-gpu-render True --simulator gpu --gpu-render-device auto --gpu-render-profile False --gpu-render-profile-interval 200 --pipe_gap 100 --bird_color yellow --pipe_color green --background day --frame_stack 4 --debug False --debug-timelimit-diagnostics False --hold-policy one_frame_then_noop --ordering-policy latest_ready",
+    "continuous_tanh_scale": 0.0,
+    "debug": false,
+    "debug_timelimit_diagnostics": false,
+    "decoder_mlp_layers": [],
+    "decorrelate_envs_on_one_worker": true,
+    "decorrelate_experience_max_seconds": 0,
+    "default_niceness": 0,
+    "device": "gpu",
+    "encoder_conv_architecture": "convnet_atari",
+    "encoder_conv_mlp_layers": [
+      512
+    ],
+    "encoder_mlp_layers": [
+      512,
+      512
+    ],
+    "env": "latency_flappy",
+    "env_fps": 10,
+    "env_frameskip": 1,
+    "env_framestack": 1,
+    "env_gpu_actions": true,
+    "env_gpu_observations": true,
+    "episode_metrics_path": "results/checkpoints/jitter_destructiveness/real_profile/teachers_latency0/flappy_teacher_fixed_l0_obs30_seed0/episode_metrics.jsonl",
+    "eval_deterministic": true,
+    "eval_env_frameskip": 1,
+    "eval_episodes": 100,
+    "eval_latency_raw_frame_values": "0,1,2,3,4,5",
+    "eval_max_steps": 7200,
+    "eval_parallel_envs": 50,
+    "eval_raw_reward": false,
+    "experiment": "flappy_teacher_fixed_l0_obs30_seed0",
+    "experiment_summaries_interval": 1,
+    "exploration_loss": "entropy",
+    "exploration_loss_coeff": 0.003,
+    "fixed_latency_ms": 0.0,
+    "flush_summaries_interval": 30,
+    "force_envs_single_thread": false,
+    "frame_stack": 4,
+    "gae_lambda": 0.95,
+    "gamma": 0.99,
+    "git_hash": "unknown",
+    "git_repo_name": "not a git repository",
+    "gpu_render_batch_size": 128,
+    "gpu_render_device": "auto",
+    "gpu_render_profile": false,
+    "gpu_render_profile_interval": 200,
+    "gym_id": "FlappyBird-v0",
+    "heartbeat_interval": 20,
+    "heartbeat_reporting_interval": 180,
+    "help": false,
+    "hold_policy": "one_frame_then_noop",
+    "initial_stddev": 1.0,
+    "keep_checkpoints": 5,
+    "kl_loss_coeff": 0.0,
+    "latency_seed": null,
+    "latency_type": "fixed",
+    "learning_rate": 0.00025,
+    "load_checkpoint_kind": "latest",
+    "log_to_file": true,
+    "lr_adaptive_max": 0.01,
+    "lr_adaptive_min": 1e-06,
+    "lr_schedule": "linear_decay",
+    "lr_schedule_kl_threshold": 0.008,
+    "max_episode_steps": 0,
+    "max_grad_norm": 0.5,
+    "max_latency_ms": null,
+    "max_pending_actions": null,
+    "max_policy_lag": 400,
+    "mean_latency_ms": null,
+    "min_latency_ms": null,
+    "mode": "train",
+    "nonlinearity": "elu",
+    "normalize_input": true,
+    "normalize_input_keys": null,
+    "normalize_obs": true,
+    "normalize_returns": true,
+    "num_batches_per_epoch": 8,
+    "num_batches_to_accumulate": 2,
+    "num_envs_per_worker": 1,
+    "num_epochs": 8,
+    "num_policies": 1,
+    "num_workers": 2,
+    "obs_fps": 10,
+    "obs_resize": "84,84",
+    "obs_scale": 255.0,
+    "obs_subtract_mean": 0.0,
+    "optimizer": "adam",
+    "ordering_policy": "latest_ready",
+    "output_dir": "outputs/jitter_destructiveness/real_profile/teacher_latency0/flappy/train/teacher/fixed_l0/obs30/seed_0",
+    "pbt_mix_policies_in_one_env": true,
+    "pbt_mutation_rate": 0.15,
+    "pbt_optimize_gamma": false,
+    "pbt_period_env_steps": 5000000,
+    "pbt_perturb_max": 1.5,
+    "pbt_perturb_min": 1.1,
+    "pbt_replace_fraction": 0.3,
+    "pbt_replace_reward_gap": 0.1,
+    "pbt_replace_reward_gap_absolute": 1e-06,
+    "pbt_start_mutation": 20000000,
+    "pbt_target_objective": "true_objective",
+    "pipe_color": "green",
+    "pipe_gap": 100,
+    "pixel_format": "CHW",
+    "policy_init_gain": 1.0,
+    "policy_initialization": "orthogonal",
+    "policy_workers_per_policy": 1,
+    "ppo_clip_ratio": 0.1,
+    "ppo_clip_value": 0.2,
+    "recurrence": 1,
+    "restart_behavior": "resume",
+    "reward_clip": 1000.0,
+    "reward_scale": 1.0,
+    "rnn_num_layers": 1,
+    "rnn_size": 512,
+    "rnn_type": "gru",
+    "rollout": 128,
+    "save_best_after": 100000,
+    "save_best_every_sec": 5,
+    "save_best_metric": "reward",
+    "save_every_sec": 600,
+    "save_milestones_sec": -1,
+    "score_limit": -1,
+    "screen_size": "",
+    "seed": 42,
+    "serial_mode": false,
+    "set_workers_cpu_affinity": true,
+    "shuffle_minibatches": false,
+    "simulator": "gpu",
+    "stats_avg": 100,
+    "std_latency_ms": null,
+    "summaries_use_frameskip": true,
+    "train_dir": "results/checkpoints/jitter_destructiveness/real_profile/teachers_latency0",
+    "train_for_env_steps": 25000000,
+    "train_for_seconds": 10000000000,
+    "use_env_info_cache": false,
+    "use_gpu_render": true,
+    "use_lidar": false,
+    "use_record_episode_statistics": false,
+    "use_rnn": false,
+    "value_bootstrap": false,
+    "value_loss_coeff": 0.5,
+    "vtrace_c": 1.0,
+    "vtrace_rho": 1.0,
+    "wandb_group": "flappy-obs30-fixed_l0",
+    "wandb_job_type": "sample_factory",
+    "wandb_project": "latency-sensitive-bench",
+    "wandb_tags": [
+      "real_profile_latency0_teacher",
+      "flappy",
+      "teacher",
+      "fixed",
+      "fixed_l0",
+      "obs30",
+      "seed0"
+    ],
+    "wandb_unique_id": "flappy-obs30-fixed_l0-s0",
+    "wandb_user": null,
+    "with_pbt": false,
+    "with_vtrace": false,
+    "with_wandb": true,
+    "worker_num_splits": 1
+  },
+  "checkpoint_experiment_dir": "/home/ubuntu/codex/hist8-vla-v2-20260923/outputs/tasks/decision_history_vla_hist8_v2/data_zero/flappy/teacher_runtime",
+  "checkpoint_path": "/home/ubuntu/codex/hist8-vla-v2-20260923/outputs/tasks/decision_history_vla_hist8_v2/data_zero/flappy/teacher_runtime/best_000047376_24281088_reward_2078.738.pth",
+  "checkpoint_train_step": 47376,
+  "config_source": "/home/ubuntu/codex/hist8-vla-v2-20260923/outputs/tasks/decision_history_vla_hist8_v2/data_zero/flappy/teacher_runtime/config.json",
+  "context_image_count": 0,
+  "context_window": 1,
+  "env_fps": 10.0,
+  "env_name": "flappy",
+  "episode_return_field": "episode_raw_return",
+  "filter_preset": "none",
+  "filter_report": {
+    "accepted_episodes": 200,
+    "accepted_train_episodes": 180,
+    "accepted_val_episodes": 20,
+    "attempted_episodes": 200,
+    "preset": "none",
+    "rejected_episodes": 0,
+    "train_rows_after_row_filter": 647783,
+    "train_rows_before_row_filter": 647783,
+    "val_rows_after_row_filter": 70375,
+    "val_rows_before_row_filter": 70375
+  },
+  "frame_stack": 4,
+  "image_resolution": [
+    512,
+    288,
+    3
+  ],
+  "image_source": "env_raw_rgb",
+  "image_transform": {
+    "gamma": 1.3,
+    "ground_fraction": 0.22,
+    "history_frames": 5,
+    "image_transform": "raw_rgb",
+    "min_alpha": 35,
+    "scroll_px_per_step": 4.0
+  },
+  "image_transform_prompt_contexts": {
+    "demon_attack_ghost_trail": " You can see the recent motion of the ship, missiles, and demons in the image. Use it to make your decision.",
+    "flappy_ghost_trail": " You can see the past motion of the bird in the image. Use it to make your decision."
+  },
+  "latency_override": null,
+  "latency_source": "checkpoint",
+  "max_episodes": 200,
+  "max_steps_per_episode": 3600,
+  "max_steps_unit": "raw_frame",
+  "num_envs": 256,
+  "obs_fps": 10.0,
+  "obs_stride_raw_frames": 1,
+  "prompt_source": "sample_factory_checkpoint:base_prompt",
+  "replay_num_envs": 256,
+  "reward_field": "raw_reward",
+  "reward_semantics": "raw_reward",
+  "rows_unit": "decision_step",
+  "schema_version": 2,
+  "split_seed": 0,
+  "total_episodes": 200,
+  "total_rows": 718158,
+  "train_episodes": 180,
+  "train_rows": 647783,
+  "val_episodes": 20,
+  "val_fraction": 0.1,
+  "val_rows": 70375
+}

zero-latency/flappy/vla/starvla-qwenpi_v3-h1/flappy-qwenpi_v3-hist8-zero-scratch-s42-v2/run_metadata.json

@@ -0,0 +1,18 @@
+{
+  "training_run_id": "flappy-qwenpi_v3-hist8-zero-scratch-s42-v2",
+  "wandb_run_id": "9be1cc91",
+  "recipe_version": "hist8-v2",
+  "history_encoding": "binary_0_1",
+  "training_condition": "zero",
+  "initialization_seed": 42,
+  "vlm_lr": 3e-06,
+  "max_train_steps": 5000,
+  "stop_after_steps": null,
+  "num_workers_per_rank": 16,
+  "eval_num_workers_per_rank": 16,
+  "prefetch_factor": 4,
+  "base_model_revision": "ebb281ec70b05090aa6165b016eac8ec08e71b17",
+  "dataset_dir": "/home/ubuntu/codex/hist8-vla-v2-20260923/outputs/tasks/decision_history_vla_hist8_v2/data_zero/flappy/lerobot",
+  "dataset_manifest_sha256": "e30d2b40648584d7a51b580c2b9444b6db01b3005f4a2ad690a73e26cc4ccddf",
+  "code_snapshot_sha256": "5ee627a7c7397387c3722678c74ec2740dc260d5f9d8fb14b0199f722b9c7962"
+}

zero-latency/flappy/vla/starvla-qwenpi_v3-h1/flappy-qwenpi_v3-hist8-zero-scratch-s42-v2/task_contract.json

@@ -0,0 +1,43 @@
+{
+  "task": "flappy",
+  "model": "qwenpi_v3",
+  "action_horizon": 1,
+  "image_frames": 1,
+  "state_source": "transport",
+  "state_dim": 24,
+  "history_decisions": 8,
+  "history_encoding": "binary_0_1",
+  "history_normalization": "identity",
+  "history_reset": "all_zero",
+  "history_order": "oldest_to_newest",
+  "history_entry_fields": [
+    "admission_bit",
+    "admitted_command_non_noop_bits",
+    "last_applied_action_non_noop_bits"
+  ],
+  "history_action_bit_order": [
+    "FLAP"
+  ],
+  "action_dim": 2,
+  "action_layout": "flappy_discrete",
+  "action_labels": [
+    "NOOP",
+    "FLAP"
+  ],
+  "task_prompts": [
+    "You are playing Flappy Bird. Pass through the pipe gaps and stay alive. Choose the action: NOOP, FLAP."
+  ],
+  "env_fps": 10,
+  "obs_fps": 10,
+  "evaluation": {
+    "episodes": 100,
+    "seeds": [
+      1000000,
+      1000099
+    ],
+    "parallel_envs": 32,
+    "capacity": 1,
+    "max_raw_frames": 3600,
+    "protocol_version": "hist8-history01-par32-v3"
+  }
+}