diff --git a/latency-aware/ant/mean_rtx3090_s2222_100ep_v1/README.md b/latency-aware/ant/mean_rtx3090_s2222_100ep_v1/README.md new file mode 100644 index 0000000000000000000000000000000000000000..8c2a7566aa0fb79b2d4bfbab0c5d3c7f20b5526e --- /dev/null +++ b/latency-aware/ant/mean_rtx3090_s2222_100ep_v1/README.md @@ -0,0 +1,5 @@ +# Ant QwenOFT mean-profile demonstrations + +100 accepted unique-seed episodes, each1000steps, strict raw episode return>2500. Split90training/10validation episodes,90000/10000rows. State27D, eight native continuous torques, current224x224RGB image, horizon1. Fixed mean90.56460638563993ms from the archived RTX3090 OpenVLA/QwenOFT profile; quantized to one10Hz raw frame,100ms effective. This new leaf preserves existing OpenVLA folder naming while identifying the actual downstream model asQwenOFT. + +Only raw demonstrations are distributed, preserving the repository convention. Training-host LeRobot conversion completed; recipe/expected metadata/audit/source versions are included. Teacher completed10,006,528frames and inference weights were uploaded/readback verified; teacher best is selected after full10Mbudget. This is data/teacher preparation, not completed VLA training or benchmark acceptance. diff --git a/latency-aware/ant/mean_rtx3090_s2222_100ep_v1/demonstrations/conversion/README.md b/latency-aware/ant/mean_rtx3090_s2222_100ep_v1/demonstrations/conversion/README.md new file mode 100644 index 0000000000000000000000000000000000000000..5d7738c49a6b82ae2b43e38df5066412c5a136b8 --- /dev/null +++ b/latency-aware/ant/mean_rtx3090_s2222_100ep_v1/demonstrations/conversion/README.md @@ -0,0 +1 @@ +Use latency-sensitive-bench@c3c6a39365a151e9b7a5e215452fd64e957c2b29 plus published preparation patches. Run convert_to_lerobot.py demonstrations/raw DEST/lerobot. The converter also creates DEST/lerobot__val and DEST/_generated_mixtures. Expected metadata and audit are included; raw image/state/action values are retained. LeRobot data remain on training host, following this repository raw-only distribution. diff --git a/latency-aware/ant/mean_rtx3090_s2222_100ep_v1/demonstrations/conversion/convert_to_lerobot.py b/latency-aware/ant/mean_rtx3090_s2222_100ep_v1/demonstrations/conversion/convert_to_lerobot.py new file mode 100644 index 0000000000000000000000000000000000000000..19107b9e2c7eddca4f8958b9a9792e801b3c78c0 --- /dev/null +++ b/latency-aware/ant/mean_rtx3090_s2222_100ep_v1/demonstrations/conversion/convert_to_lerobot.py @@ -0,0 +1,4 @@ +import sys +from pathlib import Path +from latency_bench.data.exporters.starvla_lerobot import convert_canonical_rollout_to_lerobot +convert_canonical_rollout_to_lerobot(Path(sys.argv[1]), Path(sys.argv[2]), action_carrier="native") diff --git a/latency-aware/ant/mean_rtx3090_s2222_100ep_v1/demonstrations/conversion/expected/_generated_mixtures/lerobot.json b/latency-aware/ant/mean_rtx3090_s2222_100ep_v1/demonstrations/conversion/expected/_generated_mixtures/lerobot.json new file mode 100644 index 0000000000000000000000000000000000000000..bd51522594357b382e27c2ff53a23c08659303ec --- /dev/null +++ b/latency-aware/ant/mean_rtx3090_s2222_100ep_v1/demonstrations/conversion/expected/_generated_mixtures/lerobot.json @@ -0,0 +1,16 @@ +{ + "lerobot": [ + [ + "lerobot", + 1.0, + "rl_games_gymnasium" + ] + ], + "lerobot__val": [ + [ + "lerobot__val", + 1.0, + "rl_games_gymnasium" + ] + ] +} \ No newline at end of file diff --git a/latency-aware/ant/mean_rtx3090_s2222_100ep_v1/demonstrations/conversion/expected/lerobot/manifest.json b/latency-aware/ant/mean_rtx3090_s2222_100ep_v1/demonstrations/conversion/expected/lerobot/manifest.json new file mode 100644 index 0000000000000000000000000000000000000000..80d740703b9a7254977b6d52f8429658a44b019a --- /dev/null +++ b/latency-aware/ant/mean_rtx3090_s2222_100ep_v1/demonstrations/conversion/expected/lerobot/manifest.json @@ -0,0 +1,232 @@ +{ + "dataset_name": "lerobot", + "env_name": "ant_rgb_state", + "episodes": 90, + "frames": 90000, + "task_prompts": [ + "Make the Ant move forward as fast as possible without falling. Predict eight continuous torques in [-1, 1] ordered as back right hip, back right ankle, front left hip, front left ankle, front right hip, front right ankle, back left hip, and back left ankle. Current action latency is 1 raw frames (100.00 ms). The environment runs at 10 FPS and observations are emitted at 10 FPS. Choose the best next action." + ], + "format": "lerobot_v2_image_parquet", + "source": "/home/ubuntu/lzj/mean-profiling/ant/data/raw", + "integration_name": "gymnasium", + "task_name": "ant_rgb_state", + "action_layout": "gymnasium_continuous_v1", + "action_labels": [ + "back_right_hip_torque", + "back_right_ankle_torque", + "front_left_hip_torque", + "front_left_ankle_torque", + "front_right_hip_torque", + "front_right_ankle_torque", + "back_left_hip_torque", + "back_left_ankle_torque" + ], + "carrier_action_labels": [ + "back_right_hip_torque", + "back_right_ankle_torque", + "front_left_hip_torque", + "front_left_ankle_torque", + "front_right_hip_torque", + "front_right_ankle_torque", + "back_left_hip_torque", + "back_left_ankle_torque" + ], + "action_dim": 8, + "active_action_dim": 8, + "action_carrier": "native", + "rows_unit": "decision_step", + "obs_fps": 10.0, + "obs_stride_raw_frames": 1, + "uses_state": true, + "state_dim": 27, + "state_labels": [ + "torso_z", + "torso_quaternion_w", + "torso_quaternion_x", + "torso_quaternion_y", + "torso_quaternion_z", + "front_left_hip_angle", + "front_left_ankle_angle", + "front_right_hip_angle", + "front_right_ankle_angle", + "back_left_hip_angle", + "back_left_ankle_angle", + "back_right_hip_angle", + "back_right_ankle_angle", + "torso_x_velocity", + "torso_y_velocity", + "torso_z_velocity", + "torso_angular_velocity_x", + "torso_angular_velocity_y", + "torso_angular_velocity_z", + "front_left_hip_angular_velocity", + "front_left_ankle_angular_velocity", + "front_right_hip_angular_velocity", + "front_right_ankle_angular_velocity", + "back_left_hip_angular_velocity", + "back_left_ankle_angular_velocity", + "back_right_hip_angular_velocity", + "back_right_ankle_angular_velocity" + ], + "state_normalization": { + "type": "min_max", + "min": [ + 0.24513374269008636, + -0.8805694580078125, + -0.33691075444221497, + -0.4456058740615845, + -0.9112266302108765, + -0.6140146255493164, + -0.09917011857032776, + -0.6453427076339722, + -1.3009648323059082, + -0.6563361883163452, + -1.2990498542785645, + -0.6459031105041504, + -0.09470116347074509, + -1.9981460571289062, + -2.8870370388031006, + -2.5440866947174072, + -5.5593695640563965, + -6.843714237213135, + -5.608675003051758, + -10.695633888244629, + -9.341012001037598, + -13.44798469543457, + -11.152595520019531, + -14.51861572265625, + -11.078145980834961, + -12.814929008483887, + -5.827189922332764 + ], + "max": [ + 0.9843788743019104, + 1.098469853401184, + 0.557304859161377, + 0.9673683047294617, + 0.7864838242530823, + 0.6336118578910828, + 1.329100489616394, + 0.624431848526001, + 0.09998779743909836, + 0.6632993221282959, + 0.09576962888240814, + 0.5784072875976562, + 1.320813536643982, + 5.847742557525635, + 3.869502544403076, + 2.964057445526123, + 4.749230861663818, + 8.3289155960083, + 4.963846206665039, + 14.281911849975586, + 11.46115779876709, + 12.22265911102295, + 7.717322826385498, + 15.017043113708496, + 6.09257173538208, + 7.276201248168945, + 11.832711219787598 + ] + }, + "robot_type": "rl_games_gymnasium", + "latency_prompt_map_path": "/home/ubuntu/lzj/mean-profiling/ant/data/lerobot/latency_prompt_map.json", + "custom_mixtures_path": "/home/ubuntu/lzj/mean-profiling/ant/data/_generated_mixtures/lerobot.json", + "gymnasium_task": { + "action_space": { + "dtype": "float32", + "high": [ + 1.0, + 1.0, + 1.0, + 1.0, + 1.0, + 1.0, + 1.0, + 1.0 + ], + "labels": [ + "back_right_hip_torque", + "back_right_ankle_torque", + "front_left_hip_torque", + "front_left_ankle_torque", + "front_right_hip_torque", + "front_right_ankle_torque", + "back_left_hip_torque", + "back_left_ankle_torque" + ], + "low": [ + -1.0, + -1.0, + -1.0, + -1.0, + -1.0, + -1.0, + -1.0, + -1.0 + ], + "type": "box" + }, + "base_prompt": "Make the Ant move forward as fast as possible without falling. Predict eight continuous torques in [-1, 1] ordered as back right hip, back right ankle, front left hip, front left ankle, front right hip, front right ankle, back left hip, and back left ankle.", + "env_fps": 10.0, + "env_id": "LatencyBench/AntContinuous-v0", + "frame_stack": 1, + "make_kwargs": { + "base_env_id": "Ant-v4", + "base_make_kwargs": { + "exclude_current_positions_from_observation": true, + "use_contact_forces": false + }, + "render_mode": "rgb_array" + }, + "noop_action": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "obs_fps": 10.0, + "registration_imports": [ + "latency_bench.envs.gymnasium_ant" + ], + "state_space": { + "labels": [ + "torso_z", + "torso_quaternion_w", + "torso_quaternion_x", + "torso_quaternion_y", + "torso_quaternion_z", + "front_left_hip_angle", + "front_left_ankle_angle", + "front_right_hip_angle", + "front_right_ankle_angle", + "back_left_hip_angle", + "back_left_ankle_angle", + "back_right_hip_angle", + "back_right_ankle_angle", + "torso_x_velocity", + "torso_y_velocity", + "torso_z_velocity", + "torso_angular_velocity_x", + "torso_angular_velocity_y", + "torso_angular_velocity_z", + "front_left_hip_angular_velocity", + "front_left_ankle_angular_velocity", + "front_right_hip_angular_velocity", + "front_right_ankle_angular_velocity", + "back_left_hip_angular_velocity", + "back_left_ankle_angular_velocity", + "back_right_hip_angular_velocity", + "back_right_ankle_angular_velocity" + ] + }, + "task_name": "ant_rgb_state" + }, + "validation_dataset_name": "lerobot__val", + "validation_episodes": 10, + "validation_frames": 10000 +} \ No newline at end of file diff --git a/latency-aware/ant/mean_rtx3090_s2222_100ep_v1/demonstrations/conversion/expected/lerobot/meta/episodes.jsonl b/latency-aware/ant/mean_rtx3090_s2222_100ep_v1/demonstrations/conversion/expected/lerobot/meta/episodes.jsonl new file mode 100644 index 0000000000000000000000000000000000000000..670bf8a3af8d3df9905e843c38105b53f8f39d76 --- /dev/null +++ b/latency-aware/ant/mean_rtx3090_s2222_100ep_v1/demonstrations/conversion/expected/lerobot/meta/episodes.jsonl @@ -0,0 +1,90 @@ +{"episode_index": 0, "length": 1000} +{"episode_index": 1, "length": 1000} +{"episode_index": 2, "length": 1000} +{"episode_index": 3, "length": 1000} +{"episode_index": 4, "length": 1000} +{"episode_index": 5, "length": 1000} +{"episode_index": 6, "length": 1000} +{"episode_index": 7, "length": 1000} +{"episode_index": 8, "length": 1000} +{"episode_index": 9, "length": 1000} +{"episode_index": 10, "length": 1000} +{"episode_index": 11, "length": 1000} +{"episode_index": 12, "length": 1000} +{"episode_index": 13, "length": 1000} +{"episode_index": 14, "length": 1000} +{"episode_index": 15, "length": 1000} +{"episode_index": 16, "length": 1000} +{"episode_index": 17, "length": 1000} +{"episode_index": 18, "length": 1000} +{"episode_index": 19, "length": 1000} +{"episode_index": 20, "length": 1000} +{"episode_index": 21, "length": 1000} +{"episode_index": 22, "length": 1000} +{"episode_index": 23, "length": 1000} +{"episode_index": 24, "length": 1000} +{"episode_index": 25, "length": 1000} +{"episode_index": 26, "length": 1000} +{"episode_index": 27, "length": 1000} +{"episode_index": 28, "length": 1000} +{"episode_index": 29, "length": 1000} +{"episode_index": 30, "length": 1000} +{"episode_index": 31, "length": 1000} +{"episode_index": 32, "length": 1000} +{"episode_index": 33, "length": 1000} +{"episode_index": 34, "length": 1000} +{"episode_index": 35, "length": 1000} +{"episode_index": 36, "length": 1000} +{"episode_index": 37, "length": 1000} +{"episode_index": 38, "length": 1000} +{"episode_index": 39, "length": 1000} +{"episode_index": 40, "length": 1000} +{"episode_index": 41, "length": 1000} +{"episode_index": 42, "length": 1000} +{"episode_index": 43, "length": 1000} +{"episode_index": 44, "length": 1000} +{"episode_index": 45, "length": 1000} +{"episode_index": 46, "length": 1000} +{"episode_index": 47, "length": 1000} +{"episode_index": 48, "length": 1000} +{"episode_index": 49, "length": 1000} +{"episode_index": 50, "length": 1000} +{"episode_index": 51, "length": 1000} +{"episode_index": 52, "length": 1000} +{"episode_index": 53, "length": 1000} +{"episode_index": 54, "length": 1000} +{"episode_index": 55, "length": 1000} +{"episode_index": 56, "length": 1000} +{"episode_index": 57, "length": 1000} +{"episode_index": 58, "length": 1000} +{"episode_index": 59, "length": 1000} +{"episode_index": 60, "length": 1000} +{"episode_index": 61, "length": 1000} +{"episode_index": 62, "length": 1000} +{"episode_index": 63, "length": 1000} +{"episode_index": 64, "length": 1000} +{"episode_index": 65, "length": 1000} +{"episode_index": 66, "length": 1000} +{"episode_index": 67, "length": 1000} +{"episode_index": 68, "length": 1000} +{"episode_index": 69, "length": 1000} +{"episode_index": 70, "length": 1000} +{"episode_index": 71, "length": 1000} +{"episode_index": 72, "length": 1000} +{"episode_index": 73, "length": 1000} +{"episode_index": 74, "length": 1000} +{"episode_index": 75, "length": 1000} +{"episode_index": 76, "length": 1000} +{"episode_index": 77, "length": 1000} +{"episode_index": 78, "length": 1000} +{"episode_index": 79, "length": 1000} +{"episode_index": 80, "length": 1000} +{"episode_index": 81, "length": 1000} +{"episode_index": 82, "length": 1000} +{"episode_index": 83, "length": 1000} +{"episode_index": 84, "length": 1000} +{"episode_index": 85, "length": 1000} +{"episode_index": 86, "length": 1000} +{"episode_index": 87, "length": 1000} +{"episode_index": 88, "length": 1000} +{"episode_index": 89, "length": 1000} diff --git a/latency-aware/ant/mean_rtx3090_s2222_100ep_v1/demonstrations/conversion/expected/lerobot/meta/info.json b/latency-aware/ant/mean_rtx3090_s2222_100ep_v1/demonstrations/conversion/expected/lerobot/meta/info.json new file mode 100644 index 0000000000000000000000000000000000000000..95b35c832b82a5738840344b3db153287dc38ef2 --- /dev/null +++ b/latency-aware/ant/mean_rtx3090_s2222_100ep_v1/demonstrations/conversion/expected/lerobot/meta/info.json @@ -0,0 +1,111 @@ +{ + "codebase_version": "v2.0", + "fps": 10.0, + "total_episodes": 90, + "total_frames": 90000, + "total_tasks": 1, + "total_videos": 0, + "total_chunks": 1, + "chunks_size": 1000, + "data_path": "data/chunk-{episode_chunk:03d}/episode_{episode_index:06d}.parquet", + "video_path": "videos/chunk-{episode_chunk:03d}/{video_key}/episode_{episode_index:06d}.mp4", + "features": { + "observation.image": { + "dtype": "image", + "shape": [ + 224, + 224, + 3 + ], + "names": [ + "height", + "width", + "channel" + ], + "video_info": { + "video.fps": 10.0 + } + }, + "observation.state": { + "dtype": "float32", + "shape": [ + 27 + ], + "names": [ + "torso_z", + "torso_quaternion_w", + "torso_quaternion_x", + "torso_quaternion_y", + "torso_quaternion_z", + "front_left_hip_angle", + "front_left_ankle_angle", + "front_right_hip_angle", + "front_right_ankle_angle", + "back_left_hip_angle", + "back_left_ankle_angle", + "back_right_hip_angle", + "back_right_ankle_angle", + "torso_x_velocity", + "torso_y_velocity", + "torso_z_velocity", + "torso_angular_velocity_x", + "torso_angular_velocity_y", + "torso_angular_velocity_z", + "front_left_hip_angular_velocity", + "front_left_ankle_angular_velocity", + "front_right_hip_angular_velocity", + "front_right_ankle_angular_velocity", + "back_left_hip_angular_velocity", + "back_left_ankle_angular_velocity", + "back_right_hip_angular_velocity", + "back_right_ankle_angular_velocity" + ] + }, + "action": { + "dtype": "float32", + "shape": [ + 8 + ], + "names": [ + "back_right_hip_torque", + "back_right_ankle_torque", + "front_left_hip_torque", + "front_left_ankle_torque", + "front_right_hip_torque", + "front_right_ankle_torque", + "back_left_hip_torque", + "back_left_ankle_torque" + ] + }, + "timestamp": { + "dtype": "float64", + "shape": [ + 1 + ] + }, + "episode_index": { + "dtype": "int64", + "shape": [ + 1 + ] + }, + "frame_index": { + "dtype": "int64", + "shape": [ + 1 + ] + }, + "task_index": { + "dtype": "int64", + "shape": [ + 1 + ] + }, + "latency": { + "dtype": "int64", + "shape": [ + 1 + ] + } + } +} \ No newline at end of file diff --git a/latency-aware/ant/mean_rtx3090_s2222_100ep_v1/demonstrations/conversion/expected/lerobot/meta/modality.json b/latency-aware/ant/mean_rtx3090_s2222_100ep_v1/demonstrations/conversion/expected/lerobot/meta/modality.json new file mode 100644 index 0000000000000000000000000000000000000000..26637180fb30552567226d0e9122431e07d50881 --- /dev/null +++ b/latency-aware/ant/mean_rtx3090_s2222_100ep_v1/demonstrations/conversion/expected/lerobot/meta/modality.json @@ -0,0 +1,30 @@ +{ + "state": { + "game_state": { + "start": 0, + "end": 27, + "dtype": "float32", + "absolute": true, + "original_key": "observation.state" + } + }, + "action": { + "button": { + "start": 0, + "end": 8, + "dtype": "float32", + "absolute": true, + "original_key": "action" + } + }, + "video": { + "image": { + "original_key": "observation.image" + } + }, + "annotation": { + "human.task_description": { + "original_key": "task_index" + } + } +} \ No newline at end of file diff --git a/latency-aware/ant/mean_rtx3090_s2222_100ep_v1/demonstrations/conversion/expected/lerobot/meta/tasks.jsonl b/latency-aware/ant/mean_rtx3090_s2222_100ep_v1/demonstrations/conversion/expected/lerobot/meta/tasks.jsonl new file mode 100644 index 0000000000000000000000000000000000000000..38463c356bf769bdaf1836b52a906eece12bfc42 --- /dev/null +++ b/latency-aware/ant/mean_rtx3090_s2222_100ep_v1/demonstrations/conversion/expected/lerobot/meta/tasks.jsonl @@ -0,0 +1 @@ +{"task": "Make the Ant move forward as fast as possible without falling. Predict eight continuous torques in [-1, 1] ordered as back right hip, back right ankle, front left hip, front left ankle, front right hip, front right ankle, back left hip, and back left ankle. Current action latency is 1 raw frames (100.00 ms). The environment runs at 10 FPS and observations are emitted at 10 FPS. Choose the best next action.", "task_index": 0} diff --git a/latency-aware/ant/mean_rtx3090_s2222_100ep_v1/demonstrations/conversion/expected/lerobot__val/manifest.json b/latency-aware/ant/mean_rtx3090_s2222_100ep_v1/demonstrations/conversion/expected/lerobot__val/manifest.json new file mode 100644 index 0000000000000000000000000000000000000000..4409ae1f62a6f1266e79f3835900d28c8375134a --- /dev/null +++ b/latency-aware/ant/mean_rtx3090_s2222_100ep_v1/demonstrations/conversion/expected/lerobot__val/manifest.json @@ -0,0 +1,230 @@ +{ + "dataset_name": "lerobot__val", + "env_name": "ant_rgb_state", + "episodes": 10, + "frames": 10000, + "task_prompts": [ + "Make the Ant move forward as fast as possible without falling. Predict eight continuous torques in [-1, 1] ordered as back right hip, back right ankle, front left hip, front left ankle, front right hip, front right ankle, back left hip, and back left ankle. Current action latency is 1 raw frames (100.00 ms). The environment runs at 10 FPS and observations are emitted at 10 FPS. Choose the best next action." + ], + "format": "lerobot_v2_image_parquet", + "source": "/home/ubuntu/lzj/mean-profiling/ant/data/raw", + "integration_name": "gymnasium", + "task_name": "ant_rgb_state", + "action_layout": "gymnasium_continuous_v1", + "action_labels": [ + "back_right_hip_torque", + "back_right_ankle_torque", + "front_left_hip_torque", + "front_left_ankle_torque", + "front_right_hip_torque", + "front_right_ankle_torque", + "back_left_hip_torque", + "back_left_ankle_torque" + ], + "carrier_action_labels": [ + "back_right_hip_torque", + "back_right_ankle_torque", + "front_left_hip_torque", + "front_left_ankle_torque", + "front_right_hip_torque", + "front_right_ankle_torque", + "back_left_hip_torque", + "back_left_ankle_torque" + ], + "action_dim": 8, + "active_action_dim": 8, + "action_carrier": "native", + "rows_unit": "decision_step", + "obs_fps": 10.0, + "obs_stride_raw_frames": 1, + "uses_state": true, + "state_dim": 27, + "state_labels": [ + "torso_z", + "torso_quaternion_w", + "torso_quaternion_x", + "torso_quaternion_y", + "torso_quaternion_z", + "front_left_hip_angle", + "front_left_ankle_angle", + "front_right_hip_angle", + "front_right_ankle_angle", + "back_left_hip_angle", + "back_left_ankle_angle", + "back_right_hip_angle", + "back_right_ankle_angle", + "torso_x_velocity", + "torso_y_velocity", + "torso_z_velocity", + "torso_angular_velocity_x", + "torso_angular_velocity_y", + "torso_angular_velocity_z", + "front_left_hip_angular_velocity", + "front_left_ankle_angular_velocity", + "front_right_hip_angular_velocity", + "front_right_ankle_angular_velocity", + "back_left_hip_angular_velocity", + "back_left_ankle_angular_velocity", + "back_right_hip_angular_velocity", + "back_right_ankle_angular_velocity" + ], + "state_normalization": { + "type": "min_max", + "min": [ + 0.24513374269008636, + -0.8805694580078125, + -0.33691075444221497, + -0.4456058740615845, + -0.9112266302108765, + -0.6140146255493164, + -0.09917011857032776, + -0.6453427076339722, + -1.3009648323059082, + -0.6563361883163452, + -1.2990498542785645, + -0.6459031105041504, + -0.09470116347074509, + -1.9981460571289062, + -2.8870370388031006, + -2.5440866947174072, + -5.5593695640563965, + -6.843714237213135, + -5.608675003051758, + -10.695633888244629, + -9.341012001037598, + -13.44798469543457, + -11.152595520019531, + -14.51861572265625, + -11.078145980834961, + -12.814929008483887, + -5.827189922332764 + ], + "max": [ + 0.9843788743019104, + 1.098469853401184, + 0.557304859161377, + 0.9673683047294617, + 0.7864838242530823, + 0.6336118578910828, + 1.329100489616394, + 0.624431848526001, + 0.09998779743909836, + 0.6632993221282959, + 0.09576962888240814, + 0.5784072875976562, + 1.320813536643982, + 5.847742557525635, + 3.869502544403076, + 2.964057445526123, + 4.749230861663818, + 8.3289155960083, + 4.963846206665039, + 14.281911849975586, + 11.46115779876709, + 12.22265911102295, + 7.717322826385498, + 15.017043113708496, + 6.09257173538208, + 7.276201248168945, + 11.832711219787598 + ] + }, + "robot_type": "rl_games_gymnasium", + "latency_prompt_map_path": "/home/ubuntu/lzj/mean-profiling/ant/data/lerobot/latency_prompt_map.json", + "custom_mixtures_path": "/home/ubuntu/lzj/mean-profiling/ant/data/_generated_mixtures/lerobot.json", + "gymnasium_task": { + "action_space": { + "dtype": "float32", + "high": [ + 1.0, + 1.0, + 1.0, + 1.0, + 1.0, + 1.0, + 1.0, + 1.0 + ], + "labels": [ + "back_right_hip_torque", + "back_right_ankle_torque", + "front_left_hip_torque", + "front_left_ankle_torque", + "front_right_hip_torque", + "front_right_ankle_torque", + "back_left_hip_torque", + "back_left_ankle_torque" + ], + "low": [ + -1.0, + -1.0, + -1.0, + -1.0, + -1.0, + -1.0, + -1.0, + -1.0 + ], + "type": "box" + }, + "base_prompt": "Make the Ant move forward as fast as possible without falling. Predict eight continuous torques in [-1, 1] ordered as back right hip, back right ankle, front left hip, front left ankle, front right hip, front right ankle, back left hip, and back left ankle.", + "env_fps": 10.0, + "env_id": "LatencyBench/AntContinuous-v0", + "frame_stack": 1, + "make_kwargs": { + "base_env_id": "Ant-v4", + "base_make_kwargs": { + "exclude_current_positions_from_observation": true, + "use_contact_forces": false + }, + "render_mode": "rgb_array" + }, + "noop_action": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "obs_fps": 10.0, + "registration_imports": [ + "latency_bench.envs.gymnasium_ant" + ], + "state_space": { + "labels": [ + "torso_z", + "torso_quaternion_w", + "torso_quaternion_x", + "torso_quaternion_y", + "torso_quaternion_z", + "front_left_hip_angle", + "front_left_ankle_angle", + "front_right_hip_angle", + "front_right_ankle_angle", + "back_left_hip_angle", + "back_left_ankle_angle", + "back_right_hip_angle", + "back_right_ankle_angle", + "torso_x_velocity", + "torso_y_velocity", + "torso_z_velocity", + "torso_angular_velocity_x", + "torso_angular_velocity_y", + "torso_angular_velocity_z", + "front_left_hip_angular_velocity", + "front_left_ankle_angular_velocity", + "front_right_hip_angular_velocity", + "front_right_ankle_angular_velocity", + "back_left_hip_angular_velocity", + "back_left_ankle_angular_velocity", + "back_right_hip_angular_velocity", + "back_right_ankle_angular_velocity" + ] + }, + "task_name": "ant_rgb_state" + }, + "training_dataset_name": "lerobot" +} \ No newline at end of file diff --git a/latency-aware/ant/mean_rtx3090_s2222_100ep_v1/demonstrations/conversion/expected/lerobot__val/meta/episodes.jsonl b/latency-aware/ant/mean_rtx3090_s2222_100ep_v1/demonstrations/conversion/expected/lerobot__val/meta/episodes.jsonl new file mode 100644 index 0000000000000000000000000000000000000000..fd2baf5c87c044c79016bddc4932765ad273b1d2 --- /dev/null +++ b/latency-aware/ant/mean_rtx3090_s2222_100ep_v1/demonstrations/conversion/expected/lerobot__val/meta/episodes.jsonl @@ -0,0 +1,10 @@ +{"episode_index": 0, "length": 1000} +{"episode_index": 1, "length": 1000} +{"episode_index": 2, "length": 1000} +{"episode_index": 3, "length": 1000} +{"episode_index": 4, "length": 1000} +{"episode_index": 5, "length": 1000} +{"episode_index": 6, "length": 1000} +{"episode_index": 7, "length": 1000} +{"episode_index": 8, "length": 1000} +{"episode_index": 9, "length": 1000} diff --git a/latency-aware/ant/mean_rtx3090_s2222_100ep_v1/demonstrations/conversion/expected/lerobot__val/meta/info.json b/latency-aware/ant/mean_rtx3090_s2222_100ep_v1/demonstrations/conversion/expected/lerobot__val/meta/info.json new file mode 100644 index 0000000000000000000000000000000000000000..e40f8135f4c4cf34a22031fc468e8d1d79a19ba4 --- /dev/null +++ b/latency-aware/ant/mean_rtx3090_s2222_100ep_v1/demonstrations/conversion/expected/lerobot__val/meta/info.json @@ -0,0 +1,111 @@ +{ + "codebase_version": "v2.0", + "fps": 10.0, + "total_episodes": 10, + "total_frames": 10000, + "total_tasks": 1, + "total_videos": 0, + "total_chunks": 1, + "chunks_size": 1000, + "data_path": "data/chunk-{episode_chunk:03d}/episode_{episode_index:06d}.parquet", + "video_path": "videos/chunk-{episode_chunk:03d}/{video_key}/episode_{episode_index:06d}.mp4", + "features": { + "observation.image": { + "dtype": "image", + "shape": [ + 224, + 224, + 3 + ], + "names": [ + "height", + "width", + "channel" + ], + "video_info": { + "video.fps": 10.0 + } + }, + "observation.state": { + "dtype": "float32", + "shape": [ + 27 + ], + "names": [ + "torso_z", + "torso_quaternion_w", + "torso_quaternion_x", + "torso_quaternion_y", + "torso_quaternion_z", + "front_left_hip_angle", + "front_left_ankle_angle", + "front_right_hip_angle", + "front_right_ankle_angle", + "back_left_hip_angle", + "back_left_ankle_angle", + "back_right_hip_angle", + "back_right_ankle_angle", + "torso_x_velocity", + "torso_y_velocity", + "torso_z_velocity", + "torso_angular_velocity_x", + "torso_angular_velocity_y", + "torso_angular_velocity_z", + "front_left_hip_angular_velocity", + "front_left_ankle_angular_velocity", + "front_right_hip_angular_velocity", + "front_right_ankle_angular_velocity", + "back_left_hip_angular_velocity", + "back_left_ankle_angular_velocity", + "back_right_hip_angular_velocity", + "back_right_ankle_angular_velocity" + ] + }, + "action": { + "dtype": "float32", + "shape": [ + 8 + ], + "names": [ + "back_right_hip_torque", + "back_right_ankle_torque", + "front_left_hip_torque", + "front_left_ankle_torque", + "front_right_hip_torque", + "front_right_ankle_torque", + "back_left_hip_torque", + "back_left_ankle_torque" + ] + }, + "timestamp": { + "dtype": "float64", + "shape": [ + 1 + ] + }, + "episode_index": { + "dtype": "int64", + "shape": [ + 1 + ] + }, + "frame_index": { + "dtype": "int64", + "shape": [ + 1 + ] + }, + "task_index": { + "dtype": "int64", + "shape": [ + 1 + ] + }, + "latency": { + "dtype": "int64", + "shape": [ + 1 + ] + } + } +} \ No newline at end of file diff --git a/latency-aware/ant/mean_rtx3090_s2222_100ep_v1/demonstrations/conversion/expected/lerobot__val/meta/modality.json b/latency-aware/ant/mean_rtx3090_s2222_100ep_v1/demonstrations/conversion/expected/lerobot__val/meta/modality.json new file mode 100644 index 0000000000000000000000000000000000000000..26637180fb30552567226d0e9122431e07d50881 --- /dev/null +++ b/latency-aware/ant/mean_rtx3090_s2222_100ep_v1/demonstrations/conversion/expected/lerobot__val/meta/modality.json @@ -0,0 +1,30 @@ +{ + "state": { + "game_state": { + "start": 0, + "end": 27, + "dtype": "float32", + "absolute": true, + "original_key": "observation.state" + } + }, + "action": { + "button": { + "start": 0, + "end": 8, + "dtype": "float32", + "absolute": true, + "original_key": "action" + } + }, + "video": { + "image": { + "original_key": "observation.image" + } + }, + "annotation": { + "human.task_description": { + "original_key": "task_index" + } + } +} \ No newline at end of file diff --git a/latency-aware/ant/mean_rtx3090_s2222_100ep_v1/demonstrations/conversion/expected/lerobot__val/meta/tasks.jsonl b/latency-aware/ant/mean_rtx3090_s2222_100ep_v1/demonstrations/conversion/expected/lerobot__val/meta/tasks.jsonl new file mode 100644 index 0000000000000000000000000000000000000000..38463c356bf769bdaf1836b52a906eece12bfc42 --- /dev/null +++ b/latency-aware/ant/mean_rtx3090_s2222_100ep_v1/demonstrations/conversion/expected/lerobot__val/meta/tasks.jsonl @@ -0,0 +1 @@ +{"task": "Make the Ant move forward as fast as possible without falling. Predict eight continuous torques in [-1, 1] ordered as back right hip, back right ankle, front left hip, front left ankle, front right hip, front right ankle, back left hip, and back left ankle. Current action latency is 1 raw frames (100.00 ms). The environment runs at 10 FPS and observations are emitted at 10 FPS. Choose the best next action.", "task_index": 0} diff --git a/latency-aware/ant/mean_rtx3090_s2222_100ep_v1/demonstrations/raw/filter_report.json b/latency-aware/ant/mean_rtx3090_s2222_100ep_v1/demonstrations/raw/filter_report.json new file mode 100644 index 0000000000000000000000000000000000000000..f125c843cdd4ca9e25a0e17c122d106a6e6cd986 --- /dev/null +++ b/latency-aware/ant/mean_rtx3090_s2222_100ep_v1/demonstrations/raw/filter_report.json @@ -0,0 +1,17 @@ +{ + "accepted_episodes": 100, + "accepted_train_episodes": 90, + "accepted_val_episodes": 10, + "attempted_episodes": 101, + "episode_return_gate": { + "field": "episode_raw_return", + "operator": ">", + "threshold": 2500.0 + }, + "preset": "none", + "rejected_episodes": 1, + "train_rows_after_row_filter": 90000, + "train_rows_before_row_filter": 90000, + "val_rows_after_row_filter": 10000, + "val_rows_before_row_filter": 10000 +} diff --git a/latency-aware/ant/mean_rtx3090_s2222_100ep_v1/demonstrations/raw/metadata.json b/latency-aware/ant/mean_rtx3090_s2222_100ep_v1/demonstrations/raw/metadata.json new file mode 100644 index 0000000000000000000000000000000000000000..c72707a5140c0f15351da0a08e446cfa064602ab --- /dev/null +++ b/latency-aware/ant/mean_rtx3090_s2222_100ep_v1/demonstrations/raw/metadata.json @@ -0,0 +1,480 @@ +{ + "action_labels": [ + "back_right_hip_torque", + "back_right_ankle_torque", + "front_left_hip_torque", + "front_left_ankle_torque", + "front_right_hip_torque", + "front_right_ankle_torque", + "back_left_hip_torque", + "back_left_ankle_torque" + ], + "action_layout": "gymnasium_continuous_v1", + "base_prompt": "Make the Ant move forward as fast as possible without falling. Predict eight continuous torques in [-1, 1] ordered as back right hip, back right ankle, front left hip, front left ankle, front right hip, front right ankle, back left hip, and back left ankle.", + "checkpoint_config": { + "action_history_decisions": null, + "actor_critic_share_weights": true, + "actor_worker_gpus": [], + "adam_beta1": 0.9, + "adam_beta2": 0.999, + "adam_eps": 1e-06, + "adaptive_stddev": false, + "add_latency_info": false, + "algo": "APPO", + "async_rl": false, + "batch_size": 1024, + "batched_sampling": false, + "benchmark": false, + "cli_args": { + "actor_critic_share_weights": true, + "adam_beta1": 0.9, + "adam_beta2": 0.999, + "adam_eps": 1e-06, + "adaptive_stddev": false, + "add_latency_info": false, + "algo": "APPO", + "async_rl": false, + "batch_size": 1024, + "batched_sampling": false, + "continuous_tanh_scale": 0.0, + "decorrelate_envs_on_one_worker": true, + "decorrelate_experience_max_seconds": 10, + "default_niceness": 0, + "device": "gpu", + "encoder_mlp_layers": [ + 64, + 64 + ], + "env": "latency_gymnasium", + "env_fps": 10.0, + "env_framestack": 1, + "episode_metrics_path": "/home/ubuntu/lzj/mean-profiling/ant/teacher/ant_mean_3090_s2222/episode_metrics.jsonl", + "eval_deterministic": true, + "eval_episodes": 10, + "eval_max_steps": 1000, + "eval_parallel_envs": 1, + "experiment": "ant_mean_3090_s2222", + "experiment_summaries_interval": 10, + "exploration_loss": "entropy", + "exploration_loss_coeff": 0.0, + "fasttd3_actor_action_l2": 0.0, + "fasttd3_compile": true, + "fasttd3_replay_batch_size": 32768, + "fasttd3_replay_capacity": 6553600, + "fasttd3_transitions_per_update": 64, + "fasttd3_v_max": 250.0, + "fasttd3_v_min": -250.0, + "force_envs_single_thread": true, + "frame_stack": 1, + "gae_lambda": 0.95, + "gamma": 0.99, + "gym_action_space_json": "{\"dtype\": \"float32\", \"high\": [1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0], \"labels\": [\"back_right_hip_torque\", \"back_right_ankle_torque\", \"front_left_hip_torque\", \"front_left_ankle_torque\", \"front_right_hip_torque\", \"front_right_ankle_torque\", \"back_left_hip_torque\", \"back_left_ankle_torque\"], \"low\": [-1.0, -1.0, -1.0, -1.0, -1.0, -1.0, -1.0, -1.0], \"type\": \"box\"}", + "gym_base_prompt": "Make the Ant move forward as fast as possible without falling. Predict eight continuous torques in [-1, 1] ordered as back right hip, back right ankle, front left hip, front left ankle, front right hip, front right ankle, back left hip, and back left ankle.", + "gym_env_id": "LatencyBench/AntContinuous-v0", + "gym_make_kwargs_json": "{\"base_env_id\": \"Ant-v4\", \"base_make_kwargs\": {\"exclude_current_positions_from_observation\": true, \"use_contact_forces\": false}, \"render_mode\": \"rgb_array\"}", + "gym_noop_action_json": "[0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0]", + "gym_registration_imports_json": "[\"latency_bench.envs.gymnasium_ant\"]", + "gym_state_labels_json": "[\"torso_z\", \"torso_quaternion_w\", \"torso_quaternion_x\", \"torso_quaternion_y\", \"torso_quaternion_z\", \"front_left_hip_angle\", \"front_left_ankle_angle\", \"front_right_hip_angle\", \"front_right_ankle_angle\", \"back_left_hip_angle\", \"back_left_ankle_angle\", \"back_right_hip_angle\", \"back_right_ankle_angle\", \"torso_x_velocity\", \"torso_y_velocity\", \"torso_z_velocity\", \"torso_angular_velocity_x\", \"torso_angular_velocity_y\", \"torso_angular_velocity_z\", \"front_left_hip_angular_velocity\", \"front_left_ankle_angular_velocity\", \"front_right_hip_angular_velocity\", \"front_right_ankle_angular_velocity\", \"back_left_hip_angular_velocity\", \"back_left_ankle_angular_velocity\", \"back_right_hip_angular_velocity\", \"back_right_ankle_angular_velocity\"]", + "gym_task_name": "ant_rgb_state", + "hold_policy": "hold", + "initial_stddev": 1.0, + "keep_checkpoints": 3, + "kl_loss_coeff": 0.1, + "latency_profile_path": null, + "latency_profile_worker_slot": null, + "latency_seed": 0, + "latency_type": "fixed", + "learning_rate": 0.00295, + "lr_schedule_kl_threshold": 0.008, + "max_grad_norm": 3.5, + "max_policy_lag": 10000, + "mode": "train", + "nonlinearity": "tanh", + "normalize_input": true, + "normalize_returns": true, + "num_batches_per_epoch": 4, + "num_batches_to_accumulate": 2, + "num_envs_per_worker": 8, + "num_epochs": 2, + "num_policies": 1, + "num_workers": 8, + "obs_fps": 10.0, + "obs_scale": 1.0, + "obs_subtract_mean": 0.0, + "optimizer": "adam", + "ordering_policy": "issue_order_fifo", + "output_dir": "/home/ubuntu/lzj/mean-profiling/ant/teacher-log", + "policy_initialization": "torch_default", + "policy_workers_per_policy": 1, + "ppo_clip_ratio": 0.2, + "ppo_clip_value": 1.0, + "recurrence": 1, + "restart_behavior": "overwrite", + "reward_clip": 1000.0, + "reward_scale": 1.0, + "rnn_size": 512, + "rnn_type": "gru", + "rollout": 64, + "save_best_after": 100000, + "save_best_every_sec": 60, + "save_every_sec": 600, + "save_milestones_sec": -1, + "seed": 2222, + "serial_mode": false, + "set_workers_cpu_affinity": true, + "shuffle_minibatches": false, + "stats_avg": 100, + "train_dir": "/home/ubuntu/lzj/mean-profiling/ant/teacher", + "train_for_env_steps": 10000000, + "use_env_info_cache": false, + "use_rnn": false, + "value_bootstrap": true, + "value_loss_coeff": 1.3, + "wandb_unique_id": "mean3090-ant-teacher-blackwell-20260930-nice0", + "with_vtrace": false, + "with_wandb": true, + "worker_num_splits": 2 + }, + "clip_reward": null, + "command_line": "--mode train --algo APPO --env latency_gymnasium --experiment ant_profile_20260911T112232Z --train_dir /mnt/results/latency-sensitive-bench/profile-latency-new/ant/20260911T112232Z-ant/small_models --restart_behavior overwrite --device gpu --seed 2222 --episode_metrics_path /mnt/results/latency-sensitive-bench/profile-latency-new/ant/20260911T112232Z-ant/small_models/ant_profile_20260911T112232Z/episode_metrics.jsonl --train_for_env_steps 10000000 --num_workers 8 --num_envs_per_worker 8 --num_policies 1 --batch_size 1024 --rollout 64 --recurrence 1 --num_epochs 2 --num_batches_per_epoch 4 --worker_num_splits 2 --max_policy_lag 10000 --num_batches_to_accumulate 2 --policy_workers_per_policy 1 --learning_rate 0.00295 --kl_loss_coeff 0.1 --lr_schedule_kl_threshold 0.008 --nonlinearity tanh --policy_initialization torch_default --continuous_tanh_scale 0.0 --initial_stddev 1.0 --env_framestack 1 --gamma 0.99 --gae_lambda 0.95 --ppo_clip_ratio 0.2 --ppo_clip_value 1.0 --exploration_loss entropy --exploration_loss_coeff 0.0 --reward_scale 1.0 --reward_clip 1000.0 --value_loss_coeff 1.3 --max_grad_norm 3.5 --optimizer adam --adam_eps 1e-06 --adam_beta1 0.9 --adam_beta2 0.999 --obs_subtract_mean 0.0 --obs_scale 1.0 --decorrelate_experience_max_seconds 10 --default_niceness 0 --rnn_type gru --rnn_size 512 --save_every_sec 600 --keep_checkpoints 3 --save_milestones_sec -1 --save_best_every_sec 60 --save_best_after 100000 --stats_avg 100 --experiment_summaries_interval 10 --fasttd3_replay_capacity 6553600 --fasttd3_replay_batch_size 32768 --fasttd3_transitions_per_update 64 --fasttd3_v_min -250.0 --fasttd3_v_max 250.0 --fasttd3_actor_action_l2 0.0 --async_rl False --batched_sampling False --serial_mode False --use_rnn False --normalize_returns True --normalize_input True --adaptive_stddev False --shuffle_minibatches False --value_bootstrap True --with_vtrace False --decorrelate_envs_on_one_worker True --set_workers_cpu_affinity True --force_envs_single_thread True --actor_critic_share_weights True --with_wandb False --fasttd3_compile True --encoder_mlp_layers 64 64 --latency-type iid --latency-seed 0 --latency-profile-path /mnt/results/latency-sensitive-bench/profile-latency-new/ant/20260911T112232Z-ant/profile/profile.json --latency-profile-worker-slot 0 --add-latency-info False --eval-episodes 10 --eval-parallel-envs 1 --eval-max-steps 1000 --eval-deterministic True --gym-task-name ant_rgb_state --gym-env-id LatencyBench/AntContinuous-v0 --gym-make-kwargs-json {\"base_env_id\": \"Ant-v4\", \"base_make_kwargs\": {\"exclude_current_positions_from_observation\": true, \"use_contact_forces\": false}, \"render_mode\": \"rgb_array\"} --gym-registration-imports-json [\"latency_bench.envs.gymnasium_ant\"] --gym-action-space-json {\"dtype\": \"float32\", \"high\": [1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0], \"labels\": [\"back_right_hip_torque\", \"back_right_ankle_torque\", \"front_left_hip_torque\", \"front_left_ankle_torque\", \"front_right_hip_torque\", \"front_right_ankle_torque\", \"back_left_hip_torque\", \"back_left_ankle_torque\"], \"low\": [-1.0, -1.0, -1.0, -1.0, -1.0, -1.0, -1.0, -1.0], \"type\": \"box\"} --gym-noop-action-json [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0] --gym-base-prompt Make the Ant move forward as fast as possible without falling. Predict eight continuous torques in [-1, 1] ordered as back right hip, back right ankle, front left hip, front left ankle, front right hip, front right ankle, back left hip, and back left ankle. --gym-state-labels-json [\"torso_z\", \"torso_quaternion_w\", \"torso_quaternion_x\", \"torso_quaternion_y\", \"torso_quaternion_z\", \"front_left_hip_angle\", \"front_left_ankle_angle\", \"front_right_hip_angle\", \"front_right_ankle_angle\", \"back_left_hip_angle\", \"back_left_ankle_angle\", \"back_right_hip_angle\", \"back_right_ankle_angle\", \"torso_x_velocity\", \"torso_y_velocity\", \"torso_z_velocity\", \"torso_angular_velocity_x\", \"torso_angular_velocity_y\", \"torso_angular_velocity_z\", \"front_left_hip_angular_velocity\", \"front_left_ankle_angular_velocity\", \"front_right_hip_angular_velocity\", \"front_right_ankle_angular_velocity\", \"back_left_hip_angular_velocity\", \"back_left_ankle_angular_velocity\", \"back_right_hip_angular_velocity\", \"back_right_ankle_angular_velocity\"] --env-fps 10.0 --obs-fps 10.0 --frame-stack 1 --hold-policy hold --ordering-policy issue_order_fifo --use_env_info_cache False", + "continuous_tanh_scale": 0.0, + "decoder_mlp_layers": [], + "decorrelate_envs_on_one_worker": true, + "decorrelate_experience_max_seconds": 10, + "default_niceness": 0, + "device": "gpu", + "encoder_conv_architecture": "convnet_simple", + "encoder_conv_mlp_layers": [ + 512 + ], + "encoder_mlp_layers": [ + 64, + 64 + ], + "env": "latency_gymnasium", + "env_fps": 10.0, + "env_frameskip": 1, + "env_framestack": 1, + "env_gpu_actions": false, + "env_gpu_observations": true, + "episode_metrics_path": "/home/ubuntu/lzj/mean-profiling/ant/teacher/ant_mean_3090_s2222/episode_metrics.jsonl", + "eval_deterministic": true, + "eval_env_frameskip": 1, + "eval_episodes": 10, + "eval_latency_raw_frame_values": null, + "eval_max_steps": 1000, + "eval_parallel_envs": 1, + "eval_raw_reward": false, + "experiment": "ant_mean_3090_s2222", + "experiment_summaries_interval": 10, + "exploration_loss": "entropy", + "exploration_loss_coeff": 0.0, + "export_env_raw_rgb_frames": false, + "fasttd3_actor_action_l2": 0.0, + "fasttd3_compile": true, + "fasttd3_replay_batch_size": 32768, + "fasttd3_replay_capacity": 6553600, + "fasttd3_sonic_decoder_path": null, + "fasttd3_transitions_per_update": 64, + "fasttd3_v_max": 250.0, + "fasttd3_v_min": -250.0, + "fixed_latency_ms": 90.56460638563993, + "flush_summaries_interval": 30, + "force_envs_single_thread": true, + "frame_stack": 1, + "gae_lambda": 0.95, + "gamma": 0.99, + "git_hash": "unknown", + "git_repo_name": "not a git repository", + "gpu_env_batch_size": 1, + "gpu_env_device": "auto", + "gym_action_labels_json": null, + "gym_action_space_json": "{\"dtype\": \"float32\", \"high\": [1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0], \"labels\": [\"back_right_hip_torque\", \"back_right_ankle_torque\", \"front_left_hip_torque\", \"front_left_ankle_torque\", \"front_right_hip_torque\", \"front_right_ankle_torque\", \"back_left_hip_torque\", \"back_left_ankle_torque\"], \"low\": [-1.0, -1.0, -1.0, -1.0, -1.0, -1.0, -1.0, -1.0], \"type\": \"box\"}", + "gym_action_values_json": null, + "gym_base_prompt": "Make the Ant move forward as fast as possible without falling. Predict eight continuous torques in [-1, 1] ordered as back right hip, back right ankle, front left hip, front left ankle, front right hip, front right ankle, back left hip, and back left ankle.", + "gym_env_id": "LatencyBench/AntContinuous-v0", + "gym_make_kwargs_json": "{\"base_env_id\": \"Ant-v4\", \"base_make_kwargs\": {\"exclude_current_positions_from_observation\": true, \"use_contact_forces\": false}, \"render_mode\": \"rgb_array\"}", + "gym_noop_action_id": null, + "gym_noop_action_json": "[0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0]", + "gym_registration_imports_json": "[\"latency_bench.envs.gymnasium_ant\"]", + "gym_state_labels_json": "[\"torso_z\", \"torso_quaternion_w\", \"torso_quaternion_x\", \"torso_quaternion_y\", \"torso_quaternion_z\", \"front_left_hip_angle\", \"front_left_ankle_angle\", \"front_right_hip_angle\", \"front_right_ankle_angle\", \"back_left_hip_angle\", \"back_left_ankle_angle\", \"back_right_hip_angle\", \"back_right_ankle_angle\", \"torso_x_velocity\", \"torso_y_velocity\", \"torso_z_velocity\", \"torso_angular_velocity_x\", \"torso_angular_velocity_y\", \"torso_angular_velocity_z\", \"front_left_hip_angular_velocity\", \"front_left_ankle_angular_velocity\", \"front_right_hip_angular_velocity\", \"front_right_ankle_angular_velocity\", \"back_left_hip_angular_velocity\", \"back_left_ankle_angular_velocity\", \"back_right_hip_angular_velocity\", \"back_right_ankle_angular_velocity\"]", + "gym_task_name": "ant_rgb_state", + "heartbeat_interval": 20, + "heartbeat_reporting_interval": 180, + "help": false, + "hold_last_chunk_action": false, + "hold_policy": "hold", + "initial_model_path": null, + "initial_stddev": 1.0, + "keep_checkpoints": 3, + "kl_loss_coeff": 0.1, + "latency_profile_path": null, + "latency_profile_worker_slot": null, + "latency_seed": 0, + "latency_type": "fixed", + "learning_rate": 0.00295, + "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": null, + "max_grad_norm": 3.5, + "max_latency_ms": null, + "max_pending_actions": null, + "max_policy_lag": 10000, + "mean_latency_ms": null, + "min_latency_ms": null, + "mode": "train", + "nonlinearity": "tanh", + "noop_max": null, + "normalize_input": true, + "normalize_input_keys": null, + "normalize_returns": true, + "num_batches_per_epoch": 4, + "num_batches_to_accumulate": 2, + "num_envs_per_worker": 8, + "num_epochs": 2, + "num_policies": 1, + "num_workers": 8, + "obs_fps": 10.0, + "obs_scale": 1.0, + "obs_subtract_mean": 0.0, + "optimizer": "adam", + "ordering_policy": "issue_order_fifo", + "output_dir": "/home/ubuntu/lzj/mean-profiling/ant/teacher-log", + "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", + "pixel_format": "CHW", + "policy_init_gain": 1.0, + "policy_initialization": "torch_default", + "policy_workers_per_policy": 1, + "ppo": null, + "ppo_clip_ratio": 0.2, + "ppo_clip_value": 1.0, + "recurrence": 1, + "restart_behavior": "overwrite", + "reward_clip": 1000.0, + "reward_scale": 1.0, + "rnn_num_layers": 1, + "rnn_size": 512, + "rnn_type": "gru", + "rollout": 64, + "rollout_profile_seed_by_reset": false, + "save_best_after": 100000, + "save_best_every_sec": 60, + "save_best_metric": "reward", + "save_every_sec": 600, + "save_milestones_sec": -1, + "seed": 2222, + "serial_mode": false, + "set_workers_cpu_affinity": true, + "shuffle_minibatches": false, + "simulated_worker_capacity": null, + "simulator": "cpu", + "stats_avg": 100, + "std_latency_ms": null, + "summaries_use_frameskip": true, + "terminal_on_life_loss": null, + "train_dir": "/home/ubuntu/lzj/mean-profiling/ant/teacher", + "train_for_env_steps": 10000000, + "train_for_seconds": 10000000000, + "use_env_info_cache": false, + "use_record_episode_statistics": false, + "use_rnn": false, + "value_bootstrap": true, + "value_loss_coeff": 1.3, + "vtrace_c": 1.0, + "vtrace_rho": 1.0, + "wandb_dir": "/home/ubuntu/lzj/mean-profiling/ant/wandb", + "wandb_group": "mean-profiling-3090-blackwell", + "wandb_job_type": "SF", + "wandb_project": "starvla_tasks", + "wandb_tags": null, + "wandb_unique_id": "mean3090-ant-teacher-blackwell-20260930-nice0", + "wandb_user": "dongqianyu99-zhejiang-university", + "with_pbt": false, + "with_vtrace": false, + "with_wandb": true, + "worker_num_splits": 2 + }, + "checkpoint_experiment_dir": "/home/ubuntu/lzj/mean-profiling/ant/teacher-publication", + "checkpoint_path": "/home/ubuntu/lzj/mean-profiling/ant/teacher-publication/checkpoint.pth", + "checkpoint_train_step": 19472, + "config_source": "/home/ubuntu/lzj/mean-profiling/ant/teacher-publication/config.json", + "context_image_count": 0, + "context_window": 1, + "env_fps": 10.0, + "env_name": "ant_rgb_state", + "episode_return_field": "episode_raw_return", + "episode_return_gate": { + "field": "episode_raw_return", + "operator": ">", + "threshold": 2500.0 + }, + "filter_preset": "none", + "filter_report": { + "accepted_episodes": 100, + "accepted_train_episodes": 90, + "accepted_val_episodes": 10, + "attempted_episodes": 101, + "episode_return_gate": { + "field": "episode_raw_return", + "operator": ">", + "threshold": 2500.0 + }, + "preset": "none", + "rejected_episodes": 1, + "train_rows_after_row_filter": 90000, + "train_rows_before_row_filter": 90000, + "val_rows_after_row_filter": 10000, + "val_rows_before_row_filter": 10000 + }, + "frame_stack": 1, + "gymnasium_task": { + "action_space": { + "dtype": "float32", + "high": [ + 1.0, + 1.0, + 1.0, + 1.0, + 1.0, + 1.0, + 1.0, + 1.0 + ], + "labels": [ + "back_right_hip_torque", + "back_right_ankle_torque", + "front_left_hip_torque", + "front_left_ankle_torque", + "front_right_hip_torque", + "front_right_ankle_torque", + "back_left_hip_torque", + "back_left_ankle_torque" + ], + "low": [ + -1.0, + -1.0, + -1.0, + -1.0, + -1.0, + -1.0, + -1.0, + -1.0 + ], + "type": "box" + }, + "base_prompt": "Make the Ant move forward as fast as possible without falling. Predict eight continuous torques in [-1, 1] ordered as back right hip, back right ankle, front left hip, front left ankle, front right hip, front right ankle, back left hip, and back left ankle.", + "env_fps": 10.0, + "env_id": "LatencyBench/AntContinuous-v0", + "frame_stack": 1, + "make_kwargs": { + "base_env_id": "Ant-v4", + "base_make_kwargs": { + "exclude_current_positions_from_observation": true, + "use_contact_forces": false + }, + "render_mode": "rgb_array" + }, + "noop_action": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "obs_fps": 10.0, + "registration_imports": [ + "latency_bench.envs.gymnasium_ant" + ], + "state_space": { + "labels": [ + "torso_z", + "torso_quaternion_w", + "torso_quaternion_x", + "torso_quaternion_y", + "torso_quaternion_z", + "front_left_hip_angle", + "front_left_ankle_angle", + "front_right_hip_angle", + "front_right_ankle_angle", + "back_left_hip_angle", + "back_left_ankle_angle", + "back_right_hip_angle", + "back_right_ankle_angle", + "torso_x_velocity", + "torso_y_velocity", + "torso_z_velocity", + "torso_angular_velocity_x", + "torso_angular_velocity_y", + "torso_angular_velocity_z", + "front_left_hip_angular_velocity", + "front_left_ankle_angular_velocity", + "front_right_hip_angular_velocity", + "front_right_ankle_angular_velocity", + "back_left_hip_angular_velocity", + "back_left_ankle_angular_velocity", + "back_right_hip_angular_velocity", + "back_right_ankle_angular_velocity" + ] + }, + "task_name": "ant_rgb_state" + }, + "image_resolution": [ + 224, + 224, + 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." + }, + "integration_name": "gymnasium", + "latency_override": null, + "latency_source": "checkpoint", + "max_episodes": 100, + "max_steps_per_episode": 1000, + "max_steps_unit": "raw_frame", + "num_envs": 12, + "obs_fps": 10.0, + "obs_stride_raw_frames": 1, + "prompt_source": "sample_factory_checkpoint:gym_base_prompt", + "prompt_timing_suffix": " Current action latency is {latency_raw_frames} raw frames ({latency_ms:.2f} ms). The environment runs at {env_fps:g} FPS and observations are emitted at {obs_fps:g} FPS. Choose the best next action.", + "replay_num_envs": 8, + "reward_field": "raw_reward", + "reward_semantics": "raw_reward", + "rows_unit": "decision_step", + "schema_version": 2, + "split_seed": 0, + "total_episodes": 100, + "total_rows": 100000, + "train_episodes": 90, + "train_rows": 90000, + "val_episodes": 10, + "val_fraction": 0.1, + "val_rows": 10000 +} diff --git a/latency-aware/ant/mean_rtx3090_s2222_100ep_v1/demonstrations/raw/train.parquet b/latency-aware/ant/mean_rtx3090_s2222_100ep_v1/demonstrations/raw/train.parquet new file mode 100644 index 0000000000000000000000000000000000000000..8a2bf14beda9e737dc394aa024b6c53f2f5aa0fa --- /dev/null +++ b/latency-aware/ant/mean_rtx3090_s2222_100ep_v1/demonstrations/raw/train.parquet @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:edd15b43f9dddc05e629dd9c1e813a94f9520a512a7d0bb01047a166138b8373 +size 1424406745 diff --git a/latency-aware/ant/mean_rtx3090_s2222_100ep_v1/demonstrations/raw/val.parquet b/latency-aware/ant/mean_rtx3090_s2222_100ep_v1/demonstrations/raw/val.parquet new file mode 100644 index 0000000000000000000000000000000000000000..e05ff628ba6093b8454144e139d53928234e4bde --- /dev/null +++ b/latency-aware/ant/mean_rtx3090_s2222_100ep_v1/demonstrations/raw/val.parquet @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e5aa108960398ae84ff32bbe8a60d49441e8d01ee070de948dd38d04daf6edd1 +size 158269930 diff --git a/latency-aware/ant/mean_rtx3090_s2222_100ep_v1/demonstrations/validation/collection.complete.json b/latency-aware/ant/mean_rtx3090_s2222_100ep_v1/demonstrations/validation/collection.complete.json new file mode 100644 index 0000000000000000000000000000000000000000..aa87000c3e3d2444f9a06fd045e3703844bbd3e8 --- /dev/null +++ b/latency-aware/ant/mean_rtx3090_s2222_100ep_v1/demonstrations/validation/collection.complete.json @@ -0,0 +1,14 @@ +{ + "task": "ant", + "teacher_checkpoint": "/home/ubuntu/lzj/mean-profiling/ant/teacher-publication/checkpoint.pth", + "teacher_sha256": "a7899b32beb9b2cfec1bcff85b285dddc243a73f1c322fafefae599df0f67b3b", + "episodes": 100, + "train_episodes": 90, + "val_episodes": 10, + "train_rows": 90000, + "val_rows": 10000, + "fixed_latency_ms": 90.56460638563993, + "action_horizon": 1, + "unique_seeds": 100, + "state_dim": 27 +} diff --git a/latency-aware/ant/mean_rtx3090_s2222_100ep_v1/demonstrations/validation/dataset.audit.json b/latency-aware/ant/mean_rtx3090_s2222_100ep_v1/demonstrations/validation/dataset.audit.json new file mode 100644 index 0000000000000000000000000000000000000000..8221b356390c6dd3888214a5a3a8a2275e221be1 --- /dev/null +++ b/latency-aware/ant/mean_rtx3090_s2222_100ep_v1/demonstrations/validation/dataset.audit.json @@ -0,0 +1,88 @@ +{ + "task": "ant", + "episodes": 100, + "rows": 100000, + "split": [ + 90, + 10 + ], + "state_dim": 27, + "action_dim": 8, + "all_numeric_finite": true, + "each_episode_1000_contiguous_steps": true, + "unique_seeds": 100, + "split_seed_overlap": 0, + "strict_return_gt2500": true, + "episode_return_min_max_mean": { + "train": [ + 2996.093242406845, + 4104.504939198494, + 3921.6442338414427 + ], + "val": [ + 3773.778401404619, + 4109.358084142208, + 3976.1818374603986 + ] + }, + "mean_nominal_ms": 90.56460638563993, + "effective_latency_frames": 1, + "effective_latency_ms": 100.0, + "sampled_image_checks": [ + [ + "train", + 0, + [ + 224, + 224 + ], + "RGB" + ], + [ + "train", + 176, + [ + 224, + 224 + ], + "RGB" + ], + [ + "train", + 351, + [ + 224, + 224 + ], + "RGB" + ], + [ + "val", + 0, + [ + 224, + 224 + ], + "RGB" + ], + [ + "val", + 20, + [ + 224, + 224 + ], + "RGB" + ], + [ + "val", + 39, + [ + 224, + 224 + ], + "RGB" + ] + ], + "converted_episode_row_counts_verified": true +} diff --git a/latency-aware/ant/mean_rtx3090_s2222_100ep_v1/evidence/collect_sf.py b/latency-aware/ant/mean_rtx3090_s2222_100ep_v1/evidence/collect_sf.py new file mode 100644 index 0000000000000000000000000000000000000000..3ccf71f870cc1c2479dfcde9ba04553e0af49b21 --- /dev/null +++ b/latency-aware/ant/mean_rtx3090_s2222_100ep_v1/evidence/collect_sf.py @@ -0,0 +1,18 @@ +from pathlib import Path +import argparse,hashlib,json,subprocess,sys +import pyarrow.parquet as pq +from latency_bench.data.exporters.starvla_lerobot import convert_canonical_rollout_to_lerobot +ROOT=Path('/home/ubuntu/lzj/mean-profiling');REPO=Path('/home/ubuntu/lzj/latency-sensitive-bench') +BUDGETS={'flappy':(200,18000,256,256,'gpu'),'deadly_corridor':(8500,3600,16,8,'cpu'),'ant':(100,1000,12,8,'gpu')} +p=argparse.ArgumentParser();p.add_argument('task',choices=BUDGETS);p.add_argument('--finalize-only',action='store_true');args=p.parse_args() +task=args.task;out=ROOT/task;cfg=json.loads((out/'teacher.flat.json').read_text());ck=Path(json.loads((out/'teacher.upload.json').read_text())['local_checkpoint']);episodes,cap,probe,replay,device=BUDGETS[task];raw=out/'data/raw' +command=[sys.executable,str(REPO/'scripts/rollout_data/export_sf_teacher_dataset.py'),'--checkpoint-root',str(ck),'--output-dir',str(raw),'--max-episodes',str(episodes),'--max-steps-per-episode',str(cap),'--num-envs',str(probe),'--replay-num-envs',str(replay),'--device',device,'--val-fraction','0.1','--split-seed','0','--context-window','1','--filter-preset','none','--deterministic','--image-writer-workers','32' if device=='gpu' else '8'] +if task=='ant':command.extend(['--episode-return-gt','2500','--export-image-size','224,224']) +else:command.extend(['--max-attempts',str(episodes)]) +(out/'collection-command.json').write_text(json.dumps(command,indent=2)+'\n') +if not args.finalize_only: + subprocess.run(command,cwd=REPO,check=True) + convert_canonical_rollout_to_lerobot(raw,out/'data/lerobot',action_carrier='native') +meta=json.loads((raw/'metadata.json').read_text());manifest=json.loads((out/'data/lerobot/manifest.json').read_text());assert meta['total_episodes']==episodes and manifest['state_dim']=={'flappy':24,'deadly_corridor':120,'ant':27}[task] +train=set(pq.read_table(raw/'train.parquet',columns=['seed']).column('seed').unique().to_pylist());val=set(pq.read_table(raw/'val.parquet',columns=['seed']).column('seed').unique().to_pylist());assert len(train)==meta['train_episodes'] and len(val)==meta['val_episodes'] and train.isdisjoint(val) +report={'task':task,'teacher_checkpoint':str(ck),'teacher_sha256':hashlib.sha256(ck.read_bytes()).hexdigest(),'episodes':episodes,'train_episodes':meta['train_episodes'],'val_episodes':meta['val_episodes'],'train_rows':meta['train_rows'],'val_rows':meta['val_rows'],'fixed_latency_ms':cfg['fixed_latency_ms'],'action_horizon':1,'unique_seeds':len(train|val),'state_dim':manifest['state_dim']};(out/'data/collection.complete.json').write_text(json.dumps(report,indent=2)+'\n');print(json.dumps(report),flush=True) diff --git a/latency-aware/ant/mean_rtx3090_s2222_100ep_v1/evidence/collection-command.json b/latency-aware/ant/mean_rtx3090_s2222_100ep_v1/evidence/collection-command.json new file mode 100644 index 0000000000000000000000000000000000000000..3bad9f930b4e2b0c717a9b84655e8d2534489e57 --- /dev/null +++ b/latency-aware/ant/mean_rtx3090_s2222_100ep_v1/evidence/collection-command.json @@ -0,0 +1,33 @@ +[ + "/home/ubuntu/lzj/conda/envs/sf-data/bin/python", + "/home/ubuntu/lzj/latency-sensitive-bench/scripts/rollout_data/export_sf_teacher_dataset.py", + "--checkpoint-root", + "/home/ubuntu/lzj/mean-profiling/ant/teacher-publication/checkpoint.pth", + "--output-dir", + "/home/ubuntu/lzj/mean-profiling/ant/data/raw", + "--max-episodes", + "100", + "--max-steps-per-episode", + "1000", + "--num-envs", + "12", + "--replay-num-envs", + "8", + "--device", + "gpu", + "--val-fraction", + "0.1", + "--split-seed", + "0", + "--context-window", + "1", + "--filter-preset", + "none", + "--deterministic", + "--image-writer-workers", + "32", + "--episode-return-gt", + "2500", + "--export-image-size", + "224,224" +] diff --git a/latency-aware/ant/mean_rtx3090_s2222_100ep_v1/evidence/gym-language-contract.patch b/latency-aware/ant/mean_rtx3090_s2222_100ep_v1/evidence/gym-language-contract.patch new file mode 100644 index 0000000000000000000000000000000000000000..f9f3acfafddcff18c8a6120610a410e67e346605 --- /dev/null +++ b/latency-aware/ant/mean_rtx3090_s2222_100ep_v1/evidence/gym-language-contract.patch @@ -0,0 +1,17 @@ +diff --git a/examples/rl_games/train_files/data_registry/data_config.py b/examples/rl_games/train_files/data_registry/data_config.py +index 2ef8b39..6d76aeb 100644 +--- a/examples/rl_games/train_files/data_registry/data_config.py ++++ b/examples/rl_games/train_files/data_registry/data_config.py +@@ -53,10 +53,10 @@ class AtlantisDataConfig(FlappyDataConfig): + + + class GymnasiumDataConfig(FlappyDataConfig): +- pass ++ language_keys = ["annotation.human.task_description"] + + +-class GymnasiumNativeDataConfig(FlappyDataConfig): ++class GymnasiumNativeDataConfig(GymnasiumDataConfig): + """Configure native Gymnasium state and action vectors.""" + + state_keys = ["state.native"] diff --git a/latency-aware/ant/mean_rtx3090_s2222_100ep_v1/evidence/profile.source.json b/latency-aware/ant/mean_rtx3090_s2222_100ep_v1/evidence/profile.source.json new file mode 100644 index 0000000000000000000000000000000000000000..bf3078569d7a19926a7fce8cbaf4b168a532139a --- /dev/null +++ b/latency-aware/ant/mean_rtx3090_s2222_100ep_v1/evidence/profile.source.json @@ -0,0 +1,66 @@ +{ + "burst_model_path": "latency_burst_model.json", + "distribution_path": "latency_distribution.json", + "env_fps": 10, + "frame_ms": 100.0, + "gpu_class": "1x-rtx3090", + "instance_id": "instance_859cf1e47bca6046", + "latency_kind": "observation_to_action_latency", + "latency_method": "temporal", + "model_id": "qwenoft", + "n_admitted_observations": 7976, + "n_capacity_drops": 294, + "n_observation_attempts": 8270, + "per_slot_summary": { + "0": { + "admitted_count": 7976, + "mean_observation_to_action_latency_ms": 90.56460638563993, + "mean_worker_service_time_ms": 90.22241150451042, + "p95_observation_to_action_latency_ms": 92.13036412000656, + "p95_worker_service_time_ms": 91.75596672296524, + "p99_worker_service_time_ms": 110.92562991380692 + } + }, + "provenance": { + "base_config": "/workspace/tasks/20260911T023128Z-p-only4/ant/profile.yaml", + "checkpoint_kind": "best", + "model_artifact": { + "checkpoint": "checkpoints/steps_5000_pytorch_model.pt", + "model_config": "config.full.yaml", + "path_in_repo": "OpenVLA/zero-latency/ant_rgb_state_l0_return_gt3000_100ep_openvla_native_sft_5k", + "repo_id": "latency-sensitive-bench/extra-envs-checkpoints", + "source": "local" + }, + "session_ids": [ + 0, + 1, + 2, + 3, + 4 + ] + }, + "sample_model_type": "hidden_regime", + "source_run_id": "20260911T033037730561Z", + "summary": { + "frame_ms": 100.0, + "max_ms": 130.29119229316711, + "mean_effective_frames": 0.9056460638563993, + "mean_ms": 90.56460638563993, + "min_ms": 87.98952293395996, + "n_samples": 7976, + "p50_frames": 0.8970676946640015, + "p50_ms": 89.70676946640015, + "p90_frames": 0.9074710392951966, + "p90_ms": 90.74710392951965, + "p95_frames": 0.9213036412000656, + "p95_ms": 92.13036412000656, + "p99_frames": 1.1217999053001404, + "p99_ms": 112.17999053001404, + "prob_latency_gt_1_frame": 0.03686058174523571, + "prob_latency_gt_2_frames": 0.0, + "prob_latency_gt_3_frames": 0.0, + "std_ms": 4.164609685850939 + }, + "visualization_path": "latency_profile.png", + "workload_id": "ant" +} diff --git a/latency-aware/ant/mean_rtx3090_s2222_100ep_v1/evidence/starvla_lerobot.source.py b/latency-aware/ant/mean_rtx3090_s2222_100ep_v1/evidence/starvla_lerobot.source.py new file mode 100644 index 0000000000000000000000000000000000000000..fba03402d2c26da376833bead62f0b17d4fed9cd --- /dev/null +++ b/latency-aware/ant/mean_rtx3090_s2222_100ep_v1/evidence/starvla_lerobot.source.py @@ -0,0 +1,1141 @@ +"""Convert canonical rollout exports to the StarVLA LeRobot image format.""" + +from __future__ import annotations + +import io +import json +import shutil +from pathlib import Path +from typing import Any, Mapping + +import numpy as np +import pyarrow as pa +import pyarrow.parquet as pq +from PIL import Image + +from latency_bench.utils.io import write_jsonl + +from latency_bench.data.action_layouts import ( + ACTION_LAYOUT_GYMNASIUM_CONTINUOUS_V1, + ACTION_LAYOUT_GYMNASIUM_DISCRETE_V1, + ACTION_LAYOUT_G1_SONIC_ABSOLUTE_JOINT_TARGET_V1, + action_spec_for_env, +) +from latency_bench.data.filters import density_weights_clean_v1 +from latency_bench.data.state_normalization import min_max_normalize_state + +STATE_DIM = 1 +BRIDGE_ACTION_DIM = 7 +LE_ROBOT_CHUNK_SIZE = 1000 +_ROBOT_TYPES_BY_ENV = { + "flappy": "rl_games_flappy", + "demon_attack": "rl_games_demon_attack", + "deadly_corridor": "rl_games_deadly_corridor", + "atlantis": "rl_games_atlantis", + "asterix": "rl_games_asterix", +} +_HISTORY_STATE_DIMS = {"flappy": 24, "demon_attack": 88, "deadly_corridor": 120, "atlantis": 56, "asterix": 136} +_DEADLY_MULTIBINARY_LABELS = [ + "MOVE_FORWARD", "MOVE_BACKWARD", "MOVE_LEFT", "MOVE_RIGHT", + "TURN_LEFT", "TURN_RIGHT", "ATTACK", +] + + +def write_starvla_modality( + dataset_dir: Path | str, + state_dim: int, + action_dim: int, + *, + action_key: str = "action", +) -> Path: + path = Path(dataset_dir) / "meta" / "modality.json" + path.parent.mkdir(parents=True, exist_ok=True) + path.write_text( + json.dumps( + { + "state": { + "state": { + "start": 0, + "end": state_dim, + "dtype": "float32", + "absolute": True, + "original_key": "observation.state", + } + }, + "action": { + action_key: { + "start": 0, + "end": action_dim, + "dtype": "float32", + "absolute": True, + "original_key": action_key, + } + }, + "video": {"ego_view": {"original_key": "observation.image"}}, + "annotation": { + "human.task_description": {"original_key": "task_index"} + }, + }, + indent=2, + ) + + "\n", + encoding="utf-8", + ) + return path + + +def _action_carrier_spec( + action_labels: list[str], + action_carrier: str, +) -> tuple[list[str], int, int, list[str]]: + if action_carrier == "native": + return action_labels, len(action_labels), STATE_DIM, ["state"] + if action_carrier != "bridge": + raise ValueError(f"Unsupported action_carrier={action_carrier!r}") + if len(action_labels) > BRIDGE_ACTION_DIM: + raise ValueError( + f"bridge action carrier supports at most {BRIDGE_ACTION_DIM} actions, " + f"got {len(action_labels)}" + ) + carrier_labels = [ + *action_labels, + *( + f"BRIDGE_PAD_{index}" + for index in range(len(action_labels), BRIDGE_ACTION_DIM) + ), + ] + state_labels = [f"BRIDGE_STATE_{index}" for index in range(BRIDGE_ACTION_DIM)] + return carrier_labels, len(action_labels), BRIDGE_ACTION_DIM, state_labels + + +def _robot_type_for_env(env_name: str, integration_name: str) -> str: + if integration_name == "gymnasium": + return "rl_games_gymnasium" + return _ROBOT_TYPES_BY_ENV[env_name] + + +def _write_starvla_handoff_files( + output_dir: Path, + val_output_dir: Path, + *, + env_name: str, + integration_name: str, + rows: list[dict[str, Any]], +) -> tuple[Path, Path]: + prompt_map = { + str(int(row["latency_raw_frames"])): { + "prompt": row["prompt"], + "latency_raw_frames": int(row["latency_raw_frames"]), + "latency_ms": float(row["latency_ms"]), + } + for row in rows + } + prompt_map_path = output_dir / "latency_prompt_map.json" + prompt_map_path.write_text(json.dumps(prompt_map, indent=2), encoding="utf-8") + + mixture_path = output_dir.parent / "_generated_mixtures" / f"{output_dir.name}.json" + mixture_path.parent.mkdir(parents=True, exist_ok=True) + robot_type = _robot_type_for_env(env_name, integration_name) + mixture_path.write_text( + json.dumps( + { + output_dir.name: [[output_dir.name, 1.0, robot_type]], + val_output_dir.name: [[val_output_dir.name, 1.0, robot_type]], + }, + indent=2, + ), + encoding="utf-8", + ) + return prompt_map_path, mixture_path + + +def _png_bytes(image: Any) -> bytes: + """Serialize a rollout image value to PNG bytes. + + Args: + image: Hugging Face image struct, PIL image, or array-like image. + + Returns: + RGB PNG bytes. + """ + if isinstance(image, Mapping): + if image["bytes"] is not None: + return bytes(image["bytes"]) + with Path(image["path"]).open("rb") as handle: + return handle.read() + if not isinstance(image, Image.Image): + image = Image.fromarray(np.asarray(image)) + image = image.convert("RGB") + buffer = io.BytesIO() + image.save(buffer, format="PNG") + return buffer.getvalue() + + +def _load_rows(raw_export_dir: Path, split: str) -> list[dict[str, Any]]: + """Load canonical rollout rows from one parquet split. + + Args: + raw_export_dir: Directory containing ``train.parquet`` and ``val.parquet``. + split: Split name to read. + + Returns: + Rows as Python dictionaries. + """ + path = raw_export_dir / f"{split}.parquet" + # Bound nested image buffers per Arrow batch; the converter still owns all rows. + return [ + row + for batch in pq.ParquetFile(path).iter_batches(batch_size=1024) + for row in batch.to_pylist() + ] + + +def _episode_groups(rows: list[dict[str, Any]]) -> list[list[dict[str, Any]]]: + """Group canonical rows by episode and order by decision step. + + Args: + rows: Canonical rollout rows from one split. + + Returns: + Episode row groups sorted by episode index. + """ + grouped: dict[int, list[dict[str, Any]]] = {} + for row in rows: + grouped.setdefault(int(row["episode_idx"]), []).append(row) + return [ + sorted(episode_rows, key=lambda row: int(row["decision_step"])) + for _, episode_rows in sorted(grouped.items()) + ] + + +def _legacy_row_action_vector(row: dict[str, Any], *, action_dim: int) -> list[float]: + if "action" in row and row["action"] is not None: + return [float(value) for value in row["action"]] + if "action_id" not in row: + raise KeyError(f"row is missing action fields; available columns: {sorted(row)}") + action_id = int(row["action_id"]) + if action_id < 0 or action_id >= action_dim: + raise ValueError( + f"action_id={action_id} is outside action range [0, {action_dim - 1}]" + ) + return [1.0 if index == action_id else 0.0 for index in range(action_dim)] + + +def _write_legacy_episode( + path: Path, + rows: list[dict[str, Any]], + *, + action_dim: int, +) -> None: + path.parent.mkdir(parents=True, exist_ok=True) + table = pa.table( + { + "observation.image": pa.array( + [{"bytes": row["image_bytes"], "path": None} for row in rows], + type=pa.struct([("bytes", pa.binary()), ("path", pa.string())]), + ), + "observation.state": pa.array( + [[0.0] * STATE_DIM for _ in rows], + type=pa.list_(pa.float32(), STATE_DIM), + ), + "action": pa.array( + [row["action"] for row in rows], + type=pa.list_(pa.float32(), action_dim), + ), + "timestamp": pa.array( + [row["timestamp"] for row in rows], type=pa.float64() + ), + "episode_index": pa.array( + [row["episode_index"] for row in rows], type=pa.int64() + ), + "frame_index": pa.array( + [row["frame_index"] for row in rows], type=pa.int64() + ), + "task_index": pa.array( + [row["task_index"] for row in rows], type=pa.int64() + ), + "done": pa.array([row["done"] for row in rows], type=pa.bool_()), + "reward": pa.array([row["reward"] for row in rows], type=pa.float32()), + "density_weight": pa.array( + [row.get("density_weight", 1.0) for row in rows], + type=pa.float32(), + ), + } + ) + pq.write_table(table, path) + + +def _write_legacy_metadata( + dataset_dir: Path, + *, + fps: float, + action_labels: list[str], + episode_lengths: list[int], + task_prompts: list[str], +) -> None: + meta_dir = dataset_dir / "meta" + meta_dir.mkdir(parents=True, exist_ok=True) + action_dim = len(action_labels) + (meta_dir / "info.json").write_text( + json.dumps( + { + "codebase_version": "v2.0", + "fps": fps, + "features": { + "observation.image": {"dtype": "image"}, + "observation.state": { + "dtype": "float32", + "shape": [STATE_DIM], + }, + "action": { + "dtype": "float32", + "shape": [action_dim], + "names": action_labels, + }, + "timestamp": {"dtype": "float64", "shape": [1]}, + "episode_index": {"dtype": "int64", "shape": [1]}, + "frame_index": {"dtype": "int64", "shape": [1]}, + "task_index": {"dtype": "int64", "shape": [1]}, + }, + }, + indent=2, + ), + encoding="utf-8", + ) + write_jsonl( + meta_dir / "episodes.jsonl", + [ + {"episode_index": index, "length": int(length)} + for index, length in enumerate(episode_lengths) + ], + ) + write_jsonl( + meta_dir / "tasks.jsonl", + [ + {"task_index": index, "task": prompt} + for index, prompt in enumerate(task_prompts) + ], + ) + + +def _convert_legacy_split( + rows: list[dict[str, Any]], + output_dir: Path, + *, + env_name: str, + action_labels: list[str], + fps: float, + reward_field: str, + density_weighting: bool, +) -> dict[str, Any]: + output_dir.mkdir(parents=True, exist_ok=True) + prompt_to_task_index: dict[str, int] = {} + task_prompts: list[str] = [] + episode_lengths: list[int] = [] + for new_episode_idx, episode_rows in enumerate(_episode_groups(rows)): + density_weights = ( + density_weights_clean_v1(env_name, episode_rows) + if density_weighting + else [1.0] * len(episode_rows) + ) + out_rows = [] + for frame_idx, row in enumerate(episode_rows): + row["latency_raw_frames"] + prompt = str(row["prompt"]) + if prompt not in prompt_to_task_index: + prompt_to_task_index[prompt] = len(task_prompts) + task_prompts.append(prompt) + out_rows.append( + { + "image_bytes": _png_bytes(row["image"]), + "action": _legacy_row_action_vector( + row, + action_dim=len(action_labels), + ), + "timestamp": frame_idx / fps, + "episode_index": new_episode_idx, + "frame_index": frame_idx, + "task_index": prompt_to_task_index[prompt], + "done": frame_idx == len(episode_rows) - 1, + "reward": float(row[reward_field]), + "density_weight": float(density_weights[frame_idx]), + } + ) + episode_lengths.append(len(out_rows)) + episode_chunk = new_episode_idx // 1000 + _write_legacy_episode( + output_dir + / f"data/chunk-{episode_chunk:03d}/episode_{new_episode_idx:06d}.parquet", + out_rows, + action_dim=len(action_labels), + ) + _write_legacy_metadata( + output_dir, + fps=fps, + action_labels=action_labels, + episode_lengths=episode_lengths, + task_prompts=task_prompts, + ) + return { + "dataset_name": output_dir.name, + "env_name": env_name, + "episodes": len(episode_lengths), + "frames": int(sum(episode_lengths)), + "task_prompts": task_prompts, + } + + +def _convert_legacy_export( + raw_export_dir: Path, + output_dir: Path, + *, + metadata: dict[str, Any], + force: bool, +) -> dict[str, Any]: + env_name = str(metadata["env_name"]) + action_spec = action_spec_for_env(env_name) + action_labels = list(metadata["action_labels"]) + reward_field = str(metadata["reward_field"]) + rows_unit = metadata["rows_unit"] + if rows_unit != "decision_step": + raise ValueError(f"rows_unit must be 'decision_step', got {rows_unit!r}") + fps = metadata["obs_fps"] + obs_stride = metadata["obs_stride_raw_frames"] + val_output_dir = output_dir.with_name(f"{output_dir.name}__val") + if output_dir.exists() and force: + shutil.rmtree(output_dir) + if val_output_dir.exists() and force: + shutil.rmtree(val_output_dir) + train_manifest = _convert_legacy_split( + _load_rows(raw_export_dir, "train"), + output_dir, + env_name=env_name, + action_labels=action_labels, + fps=fps, + reward_field=reward_field, + density_weighting=True, + ) + val_manifest = _convert_legacy_split( + _load_rows(raw_export_dir, "val"), + val_output_dir, + env_name=env_name, + action_labels=action_labels, + fps=fps, + reward_field=reward_field, + density_weighting=False, + ) + manifest = { + **train_manifest, + "format": "starvla_lerobot_v2_image_parquet", + "source": str(raw_export_dir), + "action_layout": str(metadata["action_layout"]), + "action_labels": action_labels, + "action_dim": len(action_labels), + "active_action_dim": len(action_spec.labels), + "rows_unit": rows_unit, + "obs_fps": fps, + "obs_stride_raw_frames": obs_stride, + "validation_dataset_name": val_output_dir.name, + "validation_episodes": val_manifest["episodes"], + "validation_frames": val_manifest["frames"], + } + (output_dir / "manifest.json").write_text( + json.dumps(manifest, indent=2), + encoding="utf-8", + ) + return manifest + + +def _convert_history_game_export( + raw_export_dir: Path, + output_dir: Path, + *, + metadata: dict[str, Any], + force: bool, +) -> dict[str, Any]: + """Convert decision-v1 game rollouts with their real pre-decision state.""" + env_name = metadata["env_name"] + state_dim = _HISTORY_STATE_DIMS[env_name] + action_labels = ( + _DEADLY_MULTIBINARY_LABELS + if env_name == "deadly_corridor" else list(metadata["action_labels"]) + ) + state_labels = [f"decision_history_{index}" for index in range(state_dim)] + train_rows = _load_rows(raw_export_dir, "train") + val_rows = _load_rows(raw_export_dir, "val") + for row in [*train_rows, *val_rows]: + if len(row["state"]) != state_dim: + raise ValueError(f"{env_name} decision history has {len(row['state'])} values, expected {state_dim}") + if env_name == "deadly_corridor": + action_id = row["action_id"] + turn, move, strafe, attack = ( + action_id // 18, (action_id // 6) % 3, (action_id // 2) % 3, action_id % 2 + ) + row["action"] = [ + float(move == 1), float(move == 2), float(strafe == 1), + float(strafe == 2), float(turn == 1), float(turn == 2), + float(attack == 1), + ] + first_row = (train_rows or val_rows)[0] + with Image.open(io.BytesIO(_png_bytes(first_row["image"]))) as image: + image_shape = (image.height, image.width, 3) + val_output_dir = output_dir.with_name(f"{output_dir.name}__val") + if output_dir.exists() and force: + shutil.rmtree(output_dir) + if val_output_dir.exists() and force: + shutil.rmtree(val_output_dir) + common = { + "env_name": env_name, + "integration_name": env_name, + "action_labels": action_labels, + "active_action_dim": len(action_labels), + "state_dim": state_dim, + "state_labels": state_labels, + "context_image_count": None, + "image_shape": image_shape, + "fps": metadata["obs_fps"], + "reward_field": metadata["reward_field"], + "state_normalization": None, + "include_state": True, + "density_weighting": False, + } + train_manifest = _convert_split(train_rows, output_dir, **common) + val_manifest = _convert_split(val_rows, val_output_dir, **common) + for directory in (output_dir, val_output_dir): + modality_path = directory / "meta/modality.json" + modality = json.loads(modality_path.read_text()) + modality["annotation"] = { + "human.action.task_description": {"original_key": "task_index"} + } + modality_path.write_text(json.dumps(modality, indent=2) + "\n") + prompt_map_path, mixture_path = _write_starvla_handoff_files( + output_dir, val_output_dir, env_name=env_name, + integration_name=env_name, rows=[*train_rows, *val_rows], + ) + contract = { + "format": "starvla_lerobot_v2_decision_v1", + "source": str(raw_export_dir), + "robot_type": _ROBOT_TYPES_BY_ENV[env_name], + "action_layout": "multibinary_7" if env_name == "deadly_corridor" else metadata["action_layout"], + "source_action_layout": metadata["action_layout"], + "action_labels": action_labels, + "action_dim": len(action_labels), + "active_action_dim": len(action_labels), + "state_source": "decision_v1", + "state_dim": state_dim, + "state_labels": state_labels, + "state_normalization": {"min": [0.0] * state_dim, "max": [1.0] * state_dim}, + "action_history_decisions": metadata["action_history_decisions"], + "rows_unit": metadata["rows_unit"], + "obs_fps": metadata["obs_fps"], + "obs_stride_raw_frames": metadata["obs_stride_raw_frames"], + "latency_prompt_map_path": str(prompt_map_path), + "custom_mixtures_path": str(mixture_path), + } + if env_name in {"atlantis", "asterix"}: + # The exported student image is current RGB; four-frame stacking belongs + # to the teacher and remains recorded in the raw collection metadata. + contract.update(integration_name="gymnasium", + gymnasium_task={**metadata["gymnasium_task"], "frame_stack": 1}) + manifest = {**train_manifest, **contract, + "validation_dataset_name": val_output_dir.name, + "validation_episodes": val_manifest["episodes"], + "validation_frames": val_manifest["frames"]} + (output_dir / "manifest.json").write_text(json.dumps(manifest, indent=2) + "\n") + (val_output_dir / "manifest.json").write_text( + json.dumps({**val_manifest, **contract, "training_dataset_name": output_dir.name}, indent=2) + "\n" + ) + return manifest + + +def _write_episode( + path: Path, + rows: list[dict[str, Any]], + *, + action_dim: int, + state_dim: int, + include_state: bool, +) -> None: + """Write one LeRobot episode parquet file. + + Args: + path: Destination episode parquet path. + rows: Converted LeRobot frame rows. + action_dim: Expected action vector dimension. + state_dim: Expected state vector dimension. + include_state: Whether rows carry real state vectors. + """ + path.parent.mkdir(parents=True, exist_ok=True) + states = ( + [row["state"] for row in rows] + if include_state + else [[0.0] * state_dim for _ in rows] + ) + columns = { + "observation.image": pa.array( + [{"bytes": row["image_bytes"], "path": None} for row in rows], + type=pa.struct([("bytes", pa.binary()), ("path", pa.string())]), + ), + "observation.state": pa.array( + states, + type=pa.list_(pa.float32(), state_dim), + ), + "action": pa.array( + [row["action"] for row in rows], + type=pa.list_(pa.float32(), action_dim), + ), + "timestamp": pa.array([row["timestamp"] for row in rows], type=pa.float64()), + "episode_index": pa.array( + [row["episode_index"] for row in rows], type=pa.int64() + ), + "frame_index": pa.array([row["frame_index"] for row in rows], type=pa.int64()), + "task_index": pa.array([row["task_index"] for row in rows], type=pa.int64()), + "latency": pa.array( + [row["latency_raw_frames"] for row in rows], type=pa.int64() + ), + "latency_raw_frames": pa.array( + [row["latency_raw_frames"] for row in rows], type=pa.int64() + ), + "done": pa.array([row["done"] for row in rows], type=pa.bool_()), + "reward": pa.array([row["reward"] for row in rows], type=pa.float32()), + # Per-row loss weight for KV-memory scheme-B supervision (clean_v1 event + # balancing without deleting rows). Uniform 1.0 when density is disabled. + "density_weight": pa.array( + [row["density_weight"] for row in rows], type=pa.float32() + ), + } + if rows and "issued_command_chunk" in rows[0]: + columns["issued_command_chunk"] = pa.array( + [row["issued_command_chunk"] for row in rows], + type=pa.list_(pa.list_(pa.float32(), action_dim), len(rows[0]["issued_command_chunk"])), + ) + for mask_key in ("action_prefix_mask", "action_loss_mask"): + if rows and mask_key in rows[0]: + columns[mask_key] = pa.array( + [row[mask_key] for row in rows], + type=pa.list_(pa.bool_(), len(rows[0][mask_key])), + ) + if rows and "action_id" in rows[0]: + columns["action_id"] = pa.array( + [row["action_id"] for row in rows], type=pa.int64() + ) + if rows and "context_image_bytes" in rows[0]: + image_type = pa.struct([("bytes", pa.binary()), ("path", pa.string())]) + columns["observation.context_images"] = pa.array( + [ + [{"bytes": value, "path": None} for value in row["context_image_bytes"]] + for row in rows + ], + type=pa.list_(image_type), + ) + table = pa.table(columns) + pq.write_table(table, path) + + +def _row_action_vector( + row: dict[str, Any], + *, + active_action_dim: int, + action_dim: int, +) -> list[float]: + values = [float(value) for value in row["action"]] + return [*values, *([0.0] * (action_dim - active_action_dim))] + + +def _write_metadata( + dataset_dir: Path, + *, + fps: float, + action_labels: list[str], + episode_lengths: list[int], + task_prompts: list[str], + image_shape: tuple[int, int, int], + state_dim: int, + state_labels: list[str], + context_image_count: int | None, +) -> None: + """Write minimal LeRobot metadata files for a split dataset. + + Args: + dataset_dir: LeRobot dataset split directory. + fps: Frame rate stored in metadata. + action_labels: Canonical action labels in vector order. + episode_lengths: Number of frames per converted episode. + task_prompts: Unique prompts used as LeRobot tasks. + """ + meta_dir = dataset_dir / "meta" + meta_dir.mkdir(parents=True, exist_ok=True) + action_dim = len(action_labels) + height, width, channels = image_shape + features = { + "observation.image": { + "dtype": "image", + "shape": [height, width, channels], + "names": ["height", "width", "channel"], + "video_info": {"video.fps": fps}, + }, + "observation.state": { + "dtype": "float32", + "shape": [state_dim], + "names": state_labels, + }, + "action": {"dtype": "float32", "shape": [action_dim], "names": action_labels}, + "timestamp": {"dtype": "float64", "shape": [1]}, + "episode_index": {"dtype": "int64", "shape": [1]}, + "frame_index": {"dtype": "int64", "shape": [1]}, + "task_index": {"dtype": "int64", "shape": [1]}, + "latency": {"dtype": "int64", "shape": [1]}, + } + if context_image_count is not None: + features["observation.context_images"] = { + "dtype": "image_sequence", + "shape": [context_image_count, height, width, channels], + "names": ["time", "height", "width", "channel"], + "video_info": {"video.fps": fps}, + } + (meta_dir / "info.json").write_text( + json.dumps( + { + "codebase_version": "v2.0", + "fps": fps, + "total_episodes": len(episode_lengths), + "total_frames": int(sum(episode_lengths)), + "total_tasks": len(task_prompts), + "total_videos": 0, + "total_chunks": (len(episode_lengths) + LE_ROBOT_CHUNK_SIZE - 1) + // LE_ROBOT_CHUNK_SIZE, + "chunks_size": LE_ROBOT_CHUNK_SIZE, + "data_path": "data/chunk-{episode_chunk:03d}/episode_{episode_index:06d}.parquet", + "video_path": ( + "videos/chunk-{episode_chunk:03d}/{video_key}/" + "episode_{episode_index:06d}.mp4" + ), + "features": features, + }, + indent=2, + ), + encoding="utf-8", + ) + (meta_dir / "modality.json").write_text( + json.dumps( + { + "state": { + "game_state": { + "start": 0, + "end": state_dim, + "dtype": "float32", + "absolute": True, + "original_key": "observation.state", + } + }, + "action": { + "button": { + "start": 0, + "end": action_dim, + "dtype": "float32", + "absolute": True, + "original_key": "action", + } + }, + "video": {"image": {"original_key": "observation.image"}}, + "annotation": { + "human.task_description": {"original_key": "task_index"} + }, + }, + indent=2, + ), + encoding="utf-8", + ) + write_jsonl( + meta_dir / "episodes.jsonl", + [ + {"episode_index": idx, "length": int(length)} + for idx, length in enumerate(episode_lengths) + ], + ) + write_jsonl( + meta_dir / "tasks.jsonl", + [ + {"task_index": idx, "task": prompt} + for idx, prompt in enumerate(task_prompts) + ], + ) + + +def _convert_split( + rows: list[dict[str, Any]], + output_dir: Path, + *, + env_name: str, + integration_name: str, + action_labels: list[str], + active_action_dim: int, + state_dim: int, + state_labels: list[str], + context_image_count: int | None, + image_shape: tuple[int, int, int], + fps: float, + reward_field: str, + state_normalization: dict[str, Any] | None, + include_state: bool, + density_weighting: bool = False, +) -> dict[str, Any]: + """Convert one canonical rollout split into a LeRobot dataset directory. + + Args: + rows: Canonical rollout rows from one split. + output_dir: Destination LeRobot split directory. + env_name: Canonical benchmark environment name. + action_labels: Canonical action labels in vector order. + fps: Frame rate used for timestamps. + reward_field: Source canonical reward column. + density_weighting: When True, write per-row ``density_weight`` from clean_v1's + event balancing (train split). When False, weights are uniform 1.0 (val + split keeps an unweighted held-out loss). + + Returns: + Split manifest summary. + """ + output_dir.mkdir(parents=True, exist_ok=True) + prompt_to_task_index: dict[str, int] = {} + task_prompts: list[str] = [] + episode_lengths: list[int] = [] + for new_episode_idx, episode_rows in enumerate(_episode_groups(rows)): + density_weights = ( + density_weights_clean_v1(env_name, episode_rows) + if density_weighting + and integration_name != "gymnasium" + else [1.0] * len(episode_rows) + ) + out_rows = [] + for frame_idx, row in enumerate(episode_rows): + # Enforce the canonical rollout contract; downstream must not read the legacy ``latency`` alias. + row["latency_raw_frames"] + prompt = str(row["prompt"]) + if prompt not in prompt_to_task_index: + prompt_to_task_index[prompt] = len(task_prompts) + task_prompts.append(prompt) + out_row = { + "image_bytes": _png_bytes(row["image"]), + "action": _row_action_vector( + row, + active_action_dim=active_action_dim, + action_dim=len(action_labels), + ), + "timestamp": frame_idx / fps, + "episode_index": new_episode_idx, + "frame_index": frame_idx, + "task_index": prompt_to_task_index[prompt], + "latency_raw_frames": int(row["latency_raw_frames"]), + # ``done`` is generated only for the LeRobot target format, not for canonical rollout rows. + "done": frame_idx == len(episode_rows) - 1, + "reward": float(row[reward_field]), + "density_weight": float(density_weights[frame_idx]), + } + if include_state: + state = np.asarray(row["state"], dtype=np.float32) + if state_normalization is not None: + state_min = np.asarray(state_normalization["min"], dtype=np.float32) + state_max = np.asarray(state_normalization["max"], dtype=np.float32) + state = min_max_normalize_state(state, state_min, state_max) + out_row["state"] = state.tolist() + if "issued_command_chunk" in row: + out_row["issued_command_chunk"] = row["issued_command_chunk"] + for mask_key in ("action_prefix_mask", "action_loss_mask"): + if mask_key in row: + out_row[mask_key] = row[mask_key] + if "action_id" in row: + out_row["action_id"] = int(row["action_id"]) + if "context_images" in row and row["context_images"] is not None: + out_row["context_image_bytes"] = [ + _png_bytes(image) for image in row["context_images"] + ] + out_rows.append(out_row) + episode_lengths.append(len(out_rows)) + episode_chunk = new_episode_idx // LE_ROBOT_CHUNK_SIZE + _write_episode( + output_dir + / f"data/chunk-{episode_chunk:03d}/episode_{new_episode_idx:06d}.parquet", + out_rows, + action_dim=len(action_labels), + state_dim=state_dim, + include_state=include_state, + ) + _write_metadata( + output_dir, + fps=fps, + action_labels=action_labels, + episode_lengths=episode_lengths, + task_prompts=task_prompts, + image_shape=image_shape, + state_dim=state_dim, + state_labels=state_labels, + context_image_count=context_image_count, + ) + if rows and "issued_command_chunk" in rows[0]: + info_path = output_dir / "meta/info.json" + info = json.loads(info_path.read_text()) + info["features"]["issued_command_chunk"] = { + "dtype": "float32", "shape": [len(rows[0]["issued_command_chunk"]), len(action_labels)], + "names": action_labels, + } + for mask_key in ("action_prefix_mask", "action_loss_mask"): + if mask_key in rows[0]: + info["features"][mask_key] = { + "dtype": "bool", "shape": [len(rows[0][mask_key])], "names": None, + } + info_path.write_text(json.dumps(info, indent=2) + "\n") + return { + "dataset_name": output_dir.name, + "env_name": env_name, + "episodes": len(episode_lengths), + "frames": int(sum(episode_lengths)), + "task_prompts": task_prompts, + } + + +def convert_canonical_rollout_to_lerobot( + raw_export_dir: Path | str, + output_dir: Path | str, + *, + force: bool = False, + action_carrier: str = "native", + normalize_state: bool = True, + write_starvla_handoff: bool = True, +) -> dict[str, Any]: + """Convert a canonical rollout export into train/val LeRobot datasets. + + Args: + raw_export_dir: Directory with canonical rollout parquet files and metadata. + output_dir: Destination LeRobot train dataset directory. + force: Remove existing train/val output directories before writing. + normalize_state: Apply the existing min-max state transform when the + canonical export includes a state vector. When disabled, an + explicit ``state`` field in canonical rows is copied as-is. + write_starvla_handoff: Write the existing StarVLA prompt/mixture handoff + files and manifest fields. + + Returns: + Manifest describing the converted train dataset and validation companion. + """ + raw_export_dir = Path(raw_export_dir) + output_dir = Path(output_dir) + metadata = json.loads( + (raw_export_dir / "metadata.json").read_text(encoding="utf-8") + ) + env_name = str(metadata["env_name"]) + integration_name = ( + metadata["integration_name"] if "integration_name" in metadata else env_name + ) + if metadata.get("action_history_decisions") == 8 and env_name in _HISTORY_STATE_DIMS: + return _convert_history_game_export( + raw_export_dir, output_dir, metadata=metadata, force=force + ) + if integration_name != "gymnasium": + return _convert_legacy_export( + raw_export_dir, + output_dir, + metadata=metadata, + force=force, + ) + action_layout = metadata["action_layout"] + active_action_labels = list(metadata["action_labels"]) + if action_layout not in ( + ACTION_LAYOUT_GYMNASIUM_DISCRETE_V1, + ACTION_LAYOUT_GYMNASIUM_CONTINUOUS_V1, + ACTION_LAYOUT_G1_SONIC_ABSOLUTE_JOINT_TARGET_V1, + ): + raise ValueError(f"Unsupported Gymnasium action_layout={action_layout!r}") + gymnasium_task = metadata["gymnasium_task"] + if env_name != gymnasium_task["task_name"]: + raise ValueError( + "Canonical Gymnasium env_name does not match gymnasium_task.task_name" + ) + continuous = action_layout in ( + ACTION_LAYOUT_GYMNASIUM_CONTINUOUS_V1, + ACTION_LAYOUT_G1_SONIC_ABSOLUTE_JOINT_TARGET_V1, + ) + if continuous and action_carrier == "native": + gymnasium_task = { + **gymnasium_task, + "action_space": { + key: value + for key, value in gymnasium_task["action_space"].items() + if key not in {"openvla_carrier_dim", "openvla_padding"} + }, + } + task_action_labels = ( + gymnasium_task["action_space"]["labels"] + if continuous + else gymnasium_task["action_labels"] + ) + if active_action_labels != task_action_labels: + raise ValueError( + "Canonical Gymnasium action_labels do not match gymnasium_task action-space labels" + ) + action_labels, active_action_dim, state_dim, state_labels = _action_carrier_spec( + active_action_labels, + action_carrier, + ) + train_rows = _load_rows(raw_export_dir, "train") + val_rows = _load_rows(raw_export_dir, "val") + raw_train_rows = train_rows + raw_val_rows = val_rows + admission_filtered = ( + action_layout == ACTION_LAYOUT_G1_SONIC_ABSOLUTE_JOINT_TARGET_V1 + and any("command_admitted" in row for row in [*train_rows, *val_rows]) + ) + if admission_filtered: + train_rows = [row for row in train_rows if row["command_admitted"]] + val_rows = [row for row in val_rows if row["command_admitted"]] + uses_state = "state_space" in gymnasium_task + if not normalize_state: + uses_state = any("state" in row for row in [*train_rows, *val_rows]) + if uses_state: + if "state_space" in gymnasium_task: + state_labels = list(gymnasium_task["state_space"]["labels"]) + state_dim = len(state_labels) + else: + state_row = next(row for row in [*train_rows, *val_rows] if "state" in row) + state_dim = len(state_row["state"]) + state_labels = [f"state_{index}" for index in range(state_dim)] + reward_field = str(metadata["reward_field"]) + rows_unit = metadata["rows_unit"] + if rows_unit != "decision_step": + raise ValueError(f"rows_unit must be 'decision_step', got {rows_unit!r}") + fps = metadata["obs_fps"] + obs_stride = metadata["obs_stride_raw_frames"] + val_output_dir = output_dir.with_name(f"{output_dir.name}__val") + if output_dir.exists() and force: + shutil.rmtree(output_dir) + if val_output_dir.exists() and force: + shutil.rmtree(val_output_dir) + state_normalization = None + if uses_state and normalize_state: + train_states = np.asarray([row["state"] for row in train_rows], dtype=np.float32) + state_normalization = { + "type": "min_max", + "min": train_states.min(axis=0).tolist(), + "max": train_states.max(axis=0).tolist(), + } + first_row = (raw_train_rows or raw_val_rows)[0] + image_source = first_row["image"] + with Image.open(io.BytesIO(_png_bytes(image_source))) as image: + width, height = image.size + image_shape = (height, width, 3) + context_image_count = ( + len(first_row["context_images"]) + if "context_images" in first_row and first_row["context_images"] is not None + else None + ) + train_manifest = _convert_split( + train_rows, + output_dir, + env_name=env_name, + integration_name=integration_name, + action_labels=action_labels, + active_action_dim=active_action_dim, + state_dim=state_dim, + state_labels=state_labels, + context_image_count=context_image_count, + image_shape=image_shape, + fps=fps, + reward_field=reward_field, + state_normalization=state_normalization, + include_state=uses_state, + density_weighting=True, + ) + val_manifest = _convert_split( + val_rows, + val_output_dir, + env_name=env_name, + integration_name=integration_name, + action_labels=action_labels, + active_action_dim=active_action_dim, + state_dim=state_dim, + state_labels=state_labels, + context_image_count=context_image_count, + image_shape=image_shape, + fps=fps, + reward_field=reward_field, + state_normalization=state_normalization, + include_state=uses_state, + density_weighting=False, + ) + if "label_provenance" in metadata: + train_info = json.loads((output_dir / "meta/info.json").read_text()) + for directory in (output_dir, val_output_dir): + info_path = directory / "meta/info.json" + info = json.loads(info_path.read_text()) + info["label_provenance"] = metadata["label_provenance"] + for key in ("issued_command_chunk", "action_prefix_mask", "action_loss_mask"): + if key in train_info["features"]: + info["features"][key] = train_info["features"][key] + info_path.write_text(json.dumps(info, indent=2) + "\n") + handoff_contract = { + "format": ( + "lerobot_v2_image_parquet" + if continuous + else "starvla_lerobot_v2_image_parquet" + ), + "source": str(raw_export_dir), + "integration_name": integration_name, + "task_name": env_name, + "action_layout": str(metadata["action_layout"]), + "action_labels": active_action_labels, + "carrier_action_labels": action_labels, + "action_dim": len(action_labels), + "active_action_dim": active_action_dim, + "action_carrier": action_carrier, + "rows_unit": rows_unit, + "obs_fps": fps, + "obs_stride_raw_frames": obs_stride, + "uses_state": uses_state, + "state_dim": state_dim, + "state_labels": state_labels, + "state_normalization": state_normalization, + } + if admission_filtered: + handoff_contract.update( + { + "admission_filter": "command_admitted", + "raw_train_frames": len(raw_train_rows), + "raw_validation_frames": len(raw_val_rows), + } + ) + if write_starvla_handoff: + handoff_contract["robot_type"] = _robot_type_for_env(env_name, integration_name) + prompt_map_path, mixture_path = _write_starvla_handoff_files( + output_dir, + val_output_dir, + env_name=env_name, + integration_name=integration_name, + rows=[*train_rows, *val_rows], + ) + handoff_contract.update( + { + "latency_prompt_map_path": str(prompt_map_path), + "custom_mixtures_path": str(mixture_path), + } + ) + if gymnasium_task is not None: + handoff_contract["gymnasium_task"] = gymnasium_task + manifest = { + **train_manifest, + **handoff_contract, + "validation_dataset_name": val_output_dir.name, + "validation_episodes": val_manifest["episodes"], + "validation_frames": val_manifest["frames"], + } + validation_manifest = { + **val_manifest, + **handoff_contract, + "training_dataset_name": output_dir.name, + } + (output_dir / "manifest.json").write_text( + json.dumps(manifest, indent=2), encoding="utf-8" + ) + (val_output_dir / "manifest.json").write_text( + json.dumps(validation_manifest, indent=2), encoding="utf-8" + ) + return manifest diff --git a/latency-aware/ant/mean_rtx3090_s2222_100ep_v1/files.sha256.json b/latency-aware/ant/mean_rtx3090_s2222_100ep_v1/files.sha256.json new file mode 100644 index 0000000000000000000000000000000000000000..fa98b486438af10eb7c7a181b3ebedde84891c17 --- /dev/null +++ b/latency-aware/ant/mean_rtx3090_s2222_100ep_v1/files.sha256.json @@ -0,0 +1,106 @@ +{ + "README.md": { + "bytes": 905, + "sha256": "5236cf2a83cce7e46a332b354809b867549ca1d3a5e743b57c877eefe8a56f7c" + }, + "demonstrations/conversion/README.md": { + "bytes": 412, + "sha256": "b0cc5a2a0681575c7fd23150d5b3211f65236832b4f9c715bd0b84404d35ccdf" + }, + "demonstrations/conversion/convert_to_lerobot.py": { + "bytes": 230, + "sha256": "4570ac62854befced86409de1c911b857d6a852202a3679fdfe94555ad84b3e7" + }, + "demonstrations/conversion/expected/_generated_mixtures/lerobot.json": { + "bytes": 186, + "sha256": "4478485033db57f5cfc8dd58edfb3ec3a240a22e9821b641ae0c822d7e1035a8" + }, + "demonstrations/conversion/expected/lerobot/manifest.json": { + "bytes": 6710, + "sha256": "ec82945da958d7e29d172928b136177d2b0e23612b62889f64b475446cd655c4" + }, + "demonstrations/conversion/expected/lerobot/meta/episodes.jsonl": { + "bytes": 3410, + "sha256": "57c46883b5b54ab359cfd33479813168e4a0c4b8998e3b3049546745f0915045" + }, + "demonstrations/conversion/expected/lerobot/meta/info.json": { + "bytes": 2518, + "sha256": "496e425ff5785622d03b1089ad1f23172f1c2590ef9b65752fa2997d3c24818b" + }, + "demonstrations/conversion/expected/lerobot/meta/modality.json": { + "bytes": 506, + "sha256": "745c1033cda84730d4f6282a152d7e6df575ea4a2fa7c113cd27d1b090fa3e84" + }, + "demonstrations/conversion/expected/lerobot/meta/tasks.jsonl": { + "bytes": 440, + "sha256": "4a8477b64b4737f643101744b72cbd64df41d52f779279cf713ca9c0537761e3" + }, + "demonstrations/conversion/expected/lerobot__val/manifest.json": { + "bytes": 6649, + "sha256": "5e626bd217af83f0e7238273774fa693c0020500f7000b0401505f464e197689" + }, + "demonstrations/conversion/expected/lerobot__val/meta/episodes.jsonl": { + "bytes": 370, + "sha256": "cc963fe6317bf45480ea4d71a7d2b8cab064a254b61eefc303c31c32334634ad" + }, + "demonstrations/conversion/expected/lerobot__val/meta/info.json": { + "bytes": 2518, + "sha256": "4f1418c0ebb1a7f5712d7e4f96816b1cfffb2887cdc5b5ffe1eb7e268494f003" + }, + "demonstrations/conversion/expected/lerobot__val/meta/modality.json": { + "bytes": 506, + "sha256": "745c1033cda84730d4f6282a152d7e6df575ea4a2fa7c113cd27d1b090fa3e84" + }, + "demonstrations/conversion/expected/lerobot__val/meta/tasks.jsonl": { + "bytes": 440, + "sha256": "4a8477b64b4737f643101744b72cbd64df41d52f779279cf713ca9c0537761e3" + }, + "demonstrations/raw/filter_report.json": { + "bytes": 440, + "sha256": "2e443c246b8c0cd1d944a1a3a1e7a1ab9dbd4797f01f3845fc2fe08e102aef80" + }, + "demonstrations/raw/metadata.json": { + "bytes": 22958, + "sha256": "b501a89bba2c781f8158640de66669e7e541e6b3629f76da2136a999b6ff578d" + }, + "demonstrations/raw/train.parquet": { + "bytes": 1424406745, + "sha256": "edd15b43f9dddc05e629dd9c1e813a94f9520a512a7d0bb01047a166138b8373" + }, + "demonstrations/raw/val.parquet": { + "bytes": 158269930, + "sha256": "e5aa108960398ae84ff32bbe8a60d49441e8d01ee070de948dd38d04daf6edd1" + }, + "demonstrations/validation/collection.complete.json": { + "bytes": 421, + "sha256": "3a21eb8ce207f265fee8197c317c6f93013e10dd96dc32f5a702b1b658773de5" + }, + "demonstrations/validation/dataset.audit.json": { + "bytes": 1250, + "sha256": "07deeb7e3cf86c00ecdf0cf7a4cf4e52279e1b2177bd5eb642c1b415d094582b" + }, + "evidence/collect_sf.py": { + "bytes": 2650, + "sha256": "883cdfc33382e3cdbb628124a4eb7c736ecaa6bdb3757d9fca00a21f4586d037" + }, + "evidence/collection-command.json": { + "bytes": 707, + "sha256": "378ec971f61b95b06844ee793bafef58003e4dfdaef35c45e0b8afda16a5c302" + }, + "evidence/gym-language-contract.patch": { + "bytes": 688, + "sha256": "9f5dee2c5b52d5507aff27373a8cf6def4959361d2c5264d5079b416f28a4c49" + }, + "evidence/profile.source.json": { + "bytes": 2148, + "sha256": "0e49f72ec05ac600a54e07bbca5af3143708444d606167f33fa21788a9fefe50" + }, + "evidence/starvla_lerobot.source.py": { + "bytes": 42762, + "sha256": "c3aae215af7c7f04051d7a2731a2007727f7ac3426fbc8d2aad2f6a276ae842b" + }, + "provenance.json": { + "bytes": 4168, + "sha256": "af7156185beaa8ff3e153c55c6c7b8aa426590cc309d60a82d8b55597b6ae8ee" + } +} diff --git a/latency-aware/ant/mean_rtx3090_s2222_100ep_v1/provenance.json b/latency-aware/ant/mean_rtx3090_s2222_100ep_v1/provenance.json new file mode 100644 index 0000000000000000000000000000000000000000..9a7a42c3691e0fe5b624989b5ef3e728ccda29c4 --- /dev/null +++ b/latency-aware/ant/mean_rtx3090_s2222_100ep_v1/provenance.json @@ -0,0 +1,152 @@ +{ + "task": "ant", + "downstream_model": "QwenOFT", + "path_family_alias": "OpenVLA", + "teacher": { + "repo_id": "latency-sensitive-bench/benchmark-models", + "revision": "1c2f38f16553622eef2c62ddde86036be989ea31", + "path": "latency-aware/ant/small-policy/sample-factory-qwenoft-rtx3090-mean-v1", + "local_checkpoint": "/home/ubuntu/lzj/mean-profiling/ant/teacher-publication/checkpoint.pth", + "status": "complete_verified", + "completed_env_steps": 10006528, + "selected_env_steps": 9969664, + "selected_source_checkpoint": "/home/ubuntu/lzj/mean-profiling/ant/teacher/ant_mean_3090_s2222/checkpoint_p0/best_000019472_9969664_reward_3371.900.pth", + "tensor_count": 15, + "weights_only_export": true, + "tensor_equality_verified": true, + "hash_readback_verified": true, + "url": "https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/1c2f38f16553622eef2c62ddde86036be989ea31/latency-aware/ant/small-policy/sample-factory-qwenoft-rtx3090-mean-v1", + "completed_utc": "2026-09-30T13:54:03.497277+00:00" + }, + "data": { + "task": "ant", + "teacher_checkpoint": "/home/ubuntu/lzj/mean-profiling/ant/teacher-publication/checkpoint.pth", + "teacher_sha256": "a7899b32beb9b2cfec1bcff85b285dddc243a73f1c322fafefae599df0f67b3b", + "episodes": 100, + "train_episodes": 90, + "val_episodes": 10, + "train_rows": 90000, + "val_rows": 10000, + "fixed_latency_ms": 90.56460638563993, + "action_horizon": 1, + "unique_seeds": 100, + "state_dim": 27 + }, + "runtime": { + "repo": "c3c6a39365a151e9b7a5e215452fd64e957c2b29", + "starvla": "ccca13c5177fe3d3c884b6e2de4965d916016649", + "runtime_fixes": [ + "mean-profile-preparation.patch", + "gym-language-contract.patch", + "loader-spawn-cache.patch", + "loader-spawn-test.patch" + ], + "pytorch3d": { + "revision": "33824be3cbc87a7dd1db0f6a9a9de9ac81b2d0ba", + "build": "transforms-only, no native render extension; QwenOFT uses transforms only" + }, + "decord": { + "version": "0.6.0", + "build": "official source CPU decoder CP310", + "wheel_sha256": "e193b356b1e984b4eff08d23b62e482c2c9e5037a6efdc0b1af47079ae2e4c47" + } + }, + "profile": { + "mean_ms": 90.56460638563993, + "profile": "/home/ubuntu/lzj/mean-profiling/reference/checkpoint-configs/latency-aware/ant/small-policy/sample-factory-qwenoft-rtx3090-v1/profile/profile.json", + "sha256": "0e49f72ec05ac600a54e07bbca5af3143708444d606167f33fa21788a9fefe50" + }, + "audit": { + "task": "ant", + "episodes": 100, + "rows": 100000, + "split": [ + 90, + 10 + ], + "state_dim": 27, + "action_dim": 8, + "all_numeric_finite": true, + "each_episode_1000_contiguous_steps": true, + "unique_seeds": 100, + "split_seed_overlap": 0, + "strict_return_gt2500": true, + "episode_return_min_max_mean": { + "train": [ + 2996.093242406845, + 4104.504939198494, + 3921.6442338414427 + ], + "val": [ + 3773.778401404619, + 4109.358084142208, + 3976.1818374603986 + ] + }, + "mean_nominal_ms": 90.56460638563993, + "effective_latency_frames": 1, + "effective_latency_ms": 100.0, + "sampled_image_checks": [ + [ + "train", + 0, + [ + 224, + 224 + ], + "RGB" + ], + [ + "train", + 176, + [ + 224, + 224 + ], + "RGB" + ], + [ + "train", + 351, + [ + 224, + 224 + ], + "RGB" + ], + [ + "val", + 0, + [ + 224, + 224 + ], + "RGB" + ], + [ + "val", + 20, + [ + 224, + 224 + ], + "RGB" + ], + [ + "val", + 39, + [ + 224, + 224 + ], + "RGB" + ] + ], + "converted_episode_row_counts_verified": true + }, + "condition": "fixed_mean_from_RTX3090_profile", + "raw_format": "canonical embeddedPNG parquet schema2", + "training_data_format": "converted LeRobot v2", + "vla_training_status": "not started", + "closed_loop_evaluation_status": "pending" +} diff --git a/latency-aware/deadly-corridor/mean_rtx3090_s42_8500ep_v1/README.md b/latency-aware/deadly-corridor/mean_rtx3090_s42_8500ep_v1/README.md new file mode 100644 index 0000000000000000000000000000000000000000..6ddc332d9e2370f0bac286ab017a9d58b7f4f067 --- /dev/null +++ b/latency-aware/deadly-corridor/mean_rtx3090_s42_8500ep_v1/README.md @@ -0,0 +1,5 @@ +# Deadly Corridor QwenOFT mean-profile demonstrations + +8500 unique-seed episodes,7650training/850validation,313925/34465rows. State120 binary decision-history H8, seven native binary buttons, original160x120RGB image,H1. Fixed mean73.69250777493353ms from RTX3090 profile, quantized3raw35Hzframes (85.7142857ms),8.75Hz observations. Original budget/cap3600 and16probe/8replay preserved. Native Doom environments recreated perseed; replay consistency verified after seed-reuse repair. + +Raw-only embeddedPNG Parquet distribution with original conversion recipe/expected metadata/audit/source code. Training-host LeRobot conversion complete. Teacher trained25006080frames under original25Mbudget, selectedbest22239232. This publication does not establish completed VLA training or closed-loop benchmark acceptance. diff --git a/latency-aware/deadly-corridor/mean_rtx3090_s42_8500ep_v1/demonstrations/conversion/README.md b/latency-aware/deadly-corridor/mean_rtx3090_s42_8500ep_v1/demonstrations/conversion/README.md new file mode 100644 index 0000000000000000000000000000000000000000..5d7738c49a6b82ae2b43e38df5066412c5a136b8 --- /dev/null +++ b/latency-aware/deadly-corridor/mean_rtx3090_s42_8500ep_v1/demonstrations/conversion/README.md @@ -0,0 +1 @@ +Use latency-sensitive-bench@c3c6a39365a151e9b7a5e215452fd64e957c2b29 plus published preparation patches. Run convert_to_lerobot.py demonstrations/raw DEST/lerobot. The converter also creates DEST/lerobot__val and DEST/_generated_mixtures. Expected metadata and audit are included; raw image/state/action values are retained. LeRobot data remain on training host, following this repository raw-only distribution. diff --git a/latency-aware/deadly-corridor/mean_rtx3090_s42_8500ep_v1/demonstrations/conversion/convert_to_lerobot.py b/latency-aware/deadly-corridor/mean_rtx3090_s42_8500ep_v1/demonstrations/conversion/convert_to_lerobot.py new file mode 100644 index 0000000000000000000000000000000000000000..19107b9e2c7eddca4f8958b9a9792e801b3c78c0 --- /dev/null +++ b/latency-aware/deadly-corridor/mean_rtx3090_s42_8500ep_v1/demonstrations/conversion/convert_to_lerobot.py @@ -0,0 +1,4 @@ +import sys +from pathlib import Path +from latency_bench.data.exporters.starvla_lerobot import convert_canonical_rollout_to_lerobot +convert_canonical_rollout_to_lerobot(Path(sys.argv[1]), Path(sys.argv[2]), action_carrier="native") diff --git a/latency-aware/deadly-corridor/mean_rtx3090_s42_8500ep_v1/demonstrations/conversion/expected/_generated_mixtures/lerobot.json b/latency-aware/deadly-corridor/mean_rtx3090_s42_8500ep_v1/demonstrations/conversion/expected/_generated_mixtures/lerobot.json new file mode 100644 index 0000000000000000000000000000000000000000..9b6e00303e92dedfe3ce936f1f456669fbe364af --- /dev/null +++ b/latency-aware/deadly-corridor/mean_rtx3090_s42_8500ep_v1/demonstrations/conversion/expected/_generated_mixtures/lerobot.json @@ -0,0 +1,16 @@ +{ + "lerobot": [ + [ + "lerobot", + 1.0, + "rl_games_deadly_corridor" + ] + ], + "lerobot__val": [ + [ + "lerobot__val", + 1.0, + "rl_games_deadly_corridor" + ] + ] +} \ No newline at end of file diff --git a/latency-aware/deadly-corridor/mean_rtx3090_s42_8500ep_v1/demonstrations/conversion/expected/lerobot/manifest.json b/latency-aware/deadly-corridor/mean_rtx3090_s42_8500ep_v1/demonstrations/conversion/expected/lerobot/manifest.json new file mode 100644 index 0000000000000000000000000000000000000000..c290b668f07fce6686632a8508794a471099a8db --- /dev/null +++ b/latency-aware/deadly-corridor/mean_rtx3090_s42_8500ep_v1/demonstrations/conversion/expected/lerobot/manifest.json @@ -0,0 +1,404 @@ +{ + "dataset_name": "lerobot", + "env_name": "deadly_corridor", + "episodes": 7650, + "frames": 313925, + "task_prompts": [ + "You are playing Deadly Corridor in VizDoom. Your goal is to survive the corridor, fight enemies, and reach the green armor vest at the far end. The available actions are: MOVE_FORWARD, MOVE_BACKWARD, MOVE_LEFT, MOVE_RIGHT, TURN_LEFT, TURN_RIGHT, and ATTACK. Choose the best next action from the current visual observation." + ], + "format": "starvla_lerobot_v2_decision_v1", + "source": "/home/ubuntu/lzj/mean-profiling/deadly_corridor/data/raw", + "robot_type": "rl_games_deadly_corridor", + "action_layout": "multibinary_7", + "source_action_layout": "deadly_corridor_joint_54", + "action_labels": [ + "MOVE_FORWARD", + "MOVE_BACKWARD", + "MOVE_LEFT", + "MOVE_RIGHT", + "TURN_LEFT", + "TURN_RIGHT", + "ATTACK" + ], + "action_dim": 7, + "active_action_dim": 7, + "state_source": "decision_v1", + "state_dim": 120, + "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", + "decision_history_24", + "decision_history_25", + "decision_history_26", + "decision_history_27", + "decision_history_28", + "decision_history_29", + "decision_history_30", + "decision_history_31", + "decision_history_32", + "decision_history_33", + "decision_history_34", + "decision_history_35", + "decision_history_36", + "decision_history_37", + "decision_history_38", + "decision_history_39", + "decision_history_40", + "decision_history_41", + "decision_history_42", + "decision_history_43", + "decision_history_44", + "decision_history_45", + "decision_history_46", + "decision_history_47", + "decision_history_48", + "decision_history_49", + "decision_history_50", + "decision_history_51", + "decision_history_52", + "decision_history_53", + "decision_history_54", + "decision_history_55", + "decision_history_56", + "decision_history_57", + "decision_history_58", + "decision_history_59", + "decision_history_60", + "decision_history_61", + "decision_history_62", + "decision_history_63", + "decision_history_64", + "decision_history_65", + "decision_history_66", + "decision_history_67", + "decision_history_68", + "decision_history_69", + "decision_history_70", + "decision_history_71", + "decision_history_72", + "decision_history_73", + "decision_history_74", + "decision_history_75", + "decision_history_76", + "decision_history_77", + "decision_history_78", + "decision_history_79", + "decision_history_80", + "decision_history_81", + "decision_history_82", + "decision_history_83", + "decision_history_84", + "decision_history_85", + "decision_history_86", + "decision_history_87", + "decision_history_88", + "decision_history_89", + "decision_history_90", + "decision_history_91", + "decision_history_92", + "decision_history_93", + "decision_history_94", + "decision_history_95", + "decision_history_96", + "decision_history_97", + "decision_history_98", + "decision_history_99", + "decision_history_100", + "decision_history_101", + "decision_history_102", + "decision_history_103", + "decision_history_104", + "decision_history_105", + "decision_history_106", + "decision_history_107", + "decision_history_108", + "decision_history_109", + "decision_history_110", + "decision_history_111", + "decision_history_112", + "decision_history_113", + "decision_history_114", + "decision_history_115", + "decision_history_116", + "decision_history_117", + "decision_history_118", + "decision_history_119" + ], + "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, + 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, + 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, + 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, + 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, + 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, + 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, + 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, + 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": 8.75, + "obs_stride_raw_frames": 4, + "latency_prompt_map_path": "/home/ubuntu/lzj/mean-profiling/deadly_corridor/data/lerobot/latency_prompt_map.json", + "custom_mixtures_path": "/home/ubuntu/lzj/mean-profiling/deadly_corridor/data/_generated_mixtures/lerobot.json", + "validation_dataset_name": "lerobot__val", + "validation_episodes": 850, + "validation_frames": 34465 +} diff --git a/latency-aware/deadly-corridor/mean_rtx3090_s42_8500ep_v1/demonstrations/conversion/expected/lerobot/meta/episodes.jsonl b/latency-aware/deadly-corridor/mean_rtx3090_s42_8500ep_v1/demonstrations/conversion/expected/lerobot/meta/episodes.jsonl new file mode 100644 index 0000000000000000000000000000000000000000..a099aa8dad7bb73a444b95feebc743356943e34a --- /dev/null +++ b/latency-aware/deadly-corridor/mean_rtx3090_s42_8500ep_v1/demonstrations/conversion/expected/lerobot/meta/episodes.jsonl @@ -0,0 +1,7650 @@ +{"episode_index": 0, "length": 19} +{"episode_index": 1, "length": 20} +{"episode_index": 2, "length": 44} +{"episode_index": 3, "length": 45} +{"episode_index": 4, "length": 45} +{"episode_index": 5, "length": 45} +{"episode_index": 6, "length": 45} +{"episode_index": 7, "length": 45} +{"episode_index": 8, "length": 45} +{"episode_index": 9, "length": 45} +{"episode_index": 10, "length": 45} +{"episode_index": 11, "length": 46} +{"episode_index": 12, "length": 46} +{"episode_index": 13, "length": 46} +{"episode_index": 14, "length": 47} +{"episode_index": 15, "length": 44} +{"episode_index": 16, "length": 19} +{"episode_index": 17, "length": 46} +{"episode_index": 18, "length": 43} +{"episode_index": 19, "length": 43} +{"episode_index": 20, "length": 45} +{"episode_index": 21, "length": 44} +{"episode_index": 22, "length": 45} +{"episode_index": 23, "length": 44} +{"episode_index": 24, "length": 45} +{"episode_index": 25, "length": 44} +{"episode_index": 26, "length": 46} +{"episode_index": 27, "length": 45} +{"episode_index": 28, "length": 47} +{"episode_index": 29, "length": 47} +{"episode_index": 30, "length": 45} +{"episode_index": 31, "length": 48} +{"episode_index": 32, "length": 32} +{"episode_index": 33, "length": 44} +{"episode_index": 34, "length": 44} +{"episode_index": 35, "length": 45} +{"episode_index": 36, "length": 44} +{"episode_index": 37, "length": 46} +{"episode_index": 38, "length": 43} +{"episode_index": 39, "length": 45} +{"episode_index": 40, "length": 45} +{"episode_index": 41, "length": 43} +{"episode_index": 42, "length": 45} +{"episode_index": 43, "length": 11} +{"episode_index": 44, "length": 19} +{"episode_index": 45, "length": 45} +{"episode_index": 46, "length": 45} +{"episode_index": 47, "length": 34} +{"episode_index": 48, "length": 45} +{"episode_index": 49, "length": 42} +{"episode_index": 50, "length": 45} +{"episode_index": 51, "length": 45} +{"episode_index": 52, "length": 45} +{"episode_index": 53, "length": 45} +{"episode_index": 54, "length": 45} +{"episode_index": 55, "length": 47} +{"episode_index": 56, "length": 45} +{"episode_index": 57, "length": 44} +{"episode_index": 58, "length": 45} +{"episode_index": 59, "length": 45} +{"episode_index": 60, "length": 44} +{"episode_index": 61, "length": 12} +{"episode_index": 62, "length": 45} +{"episode_index": 63, "length": 46} +{"episode_index": 64, "length": 32} +{"episode_index": 65, "length": 45} +{"episode_index": 66, "length": 44} +{"episode_index": 67, "length": 47} +{"episode_index": 68, "length": 46} +{"episode_index": 69, "length": 12} +{"episode_index": 70, "length": 45} +{"episode_index": 71, "length": 45} +{"episode_index": 72, "length": 45} +{"episode_index": 73, "length": 44} +{"episode_index": 74, "length": 49} +{"episode_index": 75, "length": 45} +{"episode_index": 76, "length": 45} +{"episode_index": 77, "length": 24} +{"episode_index": 78, "length": 31} +{"episode_index": 79, "length": 44} +{"episode_index": 80, "length": 44} +{"episode_index": 81, "length": 45} +{"episode_index": 82, "length": 45} +{"episode_index": 83, "length": 43} +{"episode_index": 84, "length": 44} +{"episode_index": 85, "length": 45} +{"episode_index": 86, "length": 45} +{"episode_index": 87, "length": 46} +{"episode_index": 88, "length": 47} +{"episode_index": 89, "length": 5} +{"episode_index": 90, "length": 45} +{"episode_index": 91, "length": 45} +{"episode_index": 92, "length": 45} +{"episode_index": 93, "length": 47} +{"episode_index": 94, "length": 10} +{"episode_index": 95, "length": 46} +{"episode_index": 96, "length": 47} +{"episode_index": 97, "length": 45} +{"episode_index": 98, "length": 46} +{"episode_index": 99, "length": 19} +{"episode_index": 100, "length": 47} +{"episode_index": 101, "length": 47} +{"episode_index": 102, "length": 45} +{"episode_index": 103, "length": 19} +{"episode_index": 104, "length": 45} +{"episode_index": 105, "length": 23} +{"episode_index": 106, "length": 11} +{"episode_index": 107, "length": 45} +{"episode_index": 108, "length": 43} +{"episode_index": 109, "length": 46} +{"episode_index": 110, "length": 45} +{"episode_index": 111, "length": 46} +{"episode_index": 112, "length": 44} +{"episode_index": 113, "length": 44} +{"episode_index": 114, "length": 46} +{"episode_index": 115, "length": 46} +{"episode_index": 116, "length": 46} +{"episode_index": 117, "length": 45} +{"episode_index": 118, "length": 23} +{"episode_index": 119, "length": 49} +{"episode_index": 120, "length": 45} +{"episode_index": 121, "length": 17} +{"episode_index": 122, "length": 44} +{"episode_index": 123, "length": 43} +{"episode_index": 124, "length": 46} +{"episode_index": 125, "length": 46} +{"episode_index": 126, "length": 43} +{"episode_index": 127, "length": 44} +{"episode_index": 128, "length": 44} +{"episode_index": 129, "length": 44} +{"episode_index": 130, "length": 45} +{"episode_index": 131, "length": 45} +{"episode_index": 132, "length": 47} +{"episode_index": 133, "length": 45} +{"episode_index": 134, "length": 44} +{"episode_index": 135, "length": 45} +{"episode_index": 136, "length": 47} +{"episode_index": 137, "length": 47} +{"episode_index": 138, "length": 46} +{"episode_index": 139, "length": 45} +{"episode_index": 140, "length": 46} +{"episode_index": 141, "length": 45} +{"episode_index": 142, "length": 45} +{"episode_index": 143, "length": 44} +{"episode_index": 144, "length": 43} +{"episode_index": 145, "length": 45} +{"episode_index": 146, "length": 45} +{"episode_index": 147, "length": 45} +{"episode_index": 148, "length": 49} +{"episode_index": 149, "length": 45} +{"episode_index": 150, "length": 45} +{"episode_index": 151, "length": 44} +{"episode_index": 152, "length": 20} +{"episode_index": 153, "length": 18} +{"episode_index": 154, "length": 19} +{"episode_index": 155, "length": 44} +{"episode_index": 156, "length": 12} +{"episode_index": 157, "length": 18} +{"episode_index": 158, "length": 45} +{"episode_index": 159, "length": 47} +{"episode_index": 160, "length": 44} +{"episode_index": 161, "length": 44} +{"episode_index": 162, "length": 46} +{"episode_index": 163, "length": 43} +{"episode_index": 164, "length": 44} +{"episode_index": 165, "length": 45} +{"episode_index": 166, "length": 44} +{"episode_index": 167, "length": 47} +{"episode_index": 168, "length": 45} +{"episode_index": 169, "length": 18} +{"episode_index": 170, "length": 44} +{"episode_index": 171, "length": 45} +{"episode_index": 172, "length": 46} +{"episode_index": 173, "length": 46} +{"episode_index": 174, "length": 45} +{"episode_index": 175, "length": 46} +{"episode_index": 176, "length": 47} +{"episode_index": 177, "length": 46} +{"episode_index": 178, "length": 46} +{"episode_index": 179, "length": 45} +{"episode_index": 180, "length": 46} +{"episode_index": 181, "length": 45} +{"episode_index": 182, "length": 46} +{"episode_index": 183, "length": 47} +{"episode_index": 184, "length": 21} +{"episode_index": 185, "length": 45} +{"episode_index": 186, "length": 45} +{"episode_index": 187, "length": 45} +{"episode_index": 188, "length": 45} +{"episode_index": 189, "length": 45} +{"episode_index": 190, "length": 45} +{"episode_index": 191, "length": 45} +{"episode_index": 192, "length": 44} +{"episode_index": 193, "length": 45} +{"episode_index": 194, "length": 46} +{"episode_index": 195, "length": 44} +{"episode_index": 196, "length": 28} +{"episode_index": 197, "length": 45} +{"episode_index": 198, "length": 45} +{"episode_index": 199, "length": 45} +{"episode_index": 200, "length": 46} +{"episode_index": 201, "length": 45} +{"episode_index": 202, "length": 45} +{"episode_index": 203, "length": 19} +{"episode_index": 204, "length": 44} +{"episode_index": 205, "length": 45} +{"episode_index": 206, "length": 45} +{"episode_index": 207, "length": 45} +{"episode_index": 208, "length": 49} +{"episode_index": 209, "length": 45} +{"episode_index": 210, "length": 45} +{"episode_index": 211, "length": 46} +{"episode_index": 212, "length": 12} +{"episode_index": 213, "length": 45} +{"episode_index": 214, "length": 45} +{"episode_index": 215, "length": 45} +{"episode_index": 216, "length": 44} +{"episode_index": 217, "length": 45} +{"episode_index": 218, "length": 19} +{"episode_index": 219, "length": 45} +{"episode_index": 220, "length": 45} +{"episode_index": 221, "length": 45} +{"episode_index": 222, "length": 46} +{"episode_index": 223, "length": 45} +{"episode_index": 224, "length": 46} +{"episode_index": 225, "length": 46} +{"episode_index": 226, "length": 43} +{"episode_index": 227, "length": 46} +{"episode_index": 228, "length": 45} +{"episode_index": 229, "length": 45} +{"episode_index": 230, "length": 43} +{"episode_index": 231, "length": 47} +{"episode_index": 232, "length": 47} +{"episode_index": 233, "length": 47} +{"episode_index": 234, "length": 30} +{"episode_index": 235, "length": 45} +{"episode_index": 236, "length": 44} +{"episode_index": 237, "length": 44} +{"episode_index": 238, "length": 18} +{"episode_index": 239, "length": 45} +{"episode_index": 240, "length": 13} +{"episode_index": 241, "length": 48} +{"episode_index": 242, "length": 45} +{"episode_index": 243, "length": 45} +{"episode_index": 244, "length": 45} +{"episode_index": 245, "length": 45} +{"episode_index": 246, "length": 45} +{"episode_index": 247, "length": 45} +{"episode_index": 248, "length": 44} +{"episode_index": 249, "length": 18} +{"episode_index": 250, "length": 11} +{"episode_index": 251, "length": 43} +{"episode_index": 252, "length": 45} +{"episode_index": 253, "length": 44} +{"episode_index": 254, "length": 49} +{"episode_index": 255, "length": 46} +{"episode_index": 256, "length": 47} +{"episode_index": 257, "length": 51} +{"episode_index": 258, "length": 46} +{"episode_index": 259, "length": 45} +{"episode_index": 260, "length": 44} +{"episode_index": 261, "length": 46} +{"episode_index": 262, "length": 47} +{"episode_index": 263, "length": 45} +{"episode_index": 264, "length": 45} +{"episode_index": 265, "length": 45} +{"episode_index": 266, "length": 45} +{"episode_index": 267, "length": 46} +{"episode_index": 268, "length": 44} +{"episode_index": 269, "length": 45} +{"episode_index": 270, "length": 46} +{"episode_index": 271, "length": 46} +{"episode_index": 272, "length": 46} +{"episode_index": 273, "length": 45} +{"episode_index": 274, "length": 44} +{"episode_index": 275, "length": 10} +{"episode_index": 276, "length": 47} +{"episode_index": 277, "length": 20} +{"episode_index": 278, "length": 46} +{"episode_index": 279, "length": 43} +{"episode_index": 280, "length": 20} +{"episode_index": 281, "length": 45} +{"episode_index": 282, "length": 26} +{"episode_index": 283, "length": 45} +{"episode_index": 284, "length": 31} +{"episode_index": 285, "length": 27} +{"episode_index": 286, "length": 45} +{"episode_index": 287, "length": 50} +{"episode_index": 288, "length": 12} +{"episode_index": 289, "length": 11} +{"episode_index": 290, "length": 47} +{"episode_index": 291, "length": 45} +{"episode_index": 292, "length": 45} +{"episode_index": 293, "length": 45} +{"episode_index": 294, "length": 44} +{"episode_index": 295, "length": 45} +{"episode_index": 296, "length": 44} +{"episode_index": 297, "length": 48} +{"episode_index": 298, "length": 45} +{"episode_index": 299, "length": 43} +{"episode_index": 300, "length": 45} +{"episode_index": 301, "length": 46} +{"episode_index": 302, "length": 43} +{"episode_index": 303, "length": 47} +{"episode_index": 304, "length": 45} +{"episode_index": 305, "length": 25} +{"episode_index": 306, "length": 18} +{"episode_index": 307, "length": 45} +{"episode_index": 308, "length": 44} +{"episode_index": 309, "length": 46} +{"episode_index": 310, "length": 33} +{"episode_index": 311, "length": 46} +{"episode_index": 312, "length": 26} +{"episode_index": 313, "length": 44} +{"episode_index": 314, "length": 46} +{"episode_index": 315, "length": 45} +{"episode_index": 316, "length": 45} +{"episode_index": 317, "length": 11} +{"episode_index": 318, "length": 45} +{"episode_index": 319, "length": 46} +{"episode_index": 320, "length": 19} +{"episode_index": 321, "length": 48} +{"episode_index": 322, "length": 12} +{"episode_index": 323, "length": 45} +{"episode_index": 324, "length": 46} +{"episode_index": 325, "length": 44} +{"episode_index": 326, "length": 46} +{"episode_index": 327, "length": 46} +{"episode_index": 328, "length": 44} +{"episode_index": 329, "length": 45} +{"episode_index": 330, "length": 45} +{"episode_index": 331, "length": 45} +{"episode_index": 332, "length": 45} +{"episode_index": 333, "length": 46} +{"episode_index": 334, "length": 44} +{"episode_index": 335, "length": 46} +{"episode_index": 336, "length": 45} +{"episode_index": 337, "length": 43} +{"episode_index": 338, "length": 46} +{"episode_index": 339, "length": 12} +{"episode_index": 340, "length": 43} +{"episode_index": 341, "length": 45} +{"episode_index": 342, "length": 46} +{"episode_index": 343, "length": 45} +{"episode_index": 344, "length": 45} +{"episode_index": 345, "length": 45} +{"episode_index": 346, "length": 45} +{"episode_index": 347, "length": 45} +{"episode_index": 348, "length": 44} +{"episode_index": 349, "length": 45} +{"episode_index": 350, "length": 44} +{"episode_index": 351, "length": 45} +{"episode_index": 352, "length": 44} +{"episode_index": 353, "length": 47} +{"episode_index": 354, "length": 44} +{"episode_index": 355, "length": 45} +{"episode_index": 356, "length": 45} +{"episode_index": 357, "length": 46} +{"episode_index": 358, "length": 46} +{"episode_index": 359, "length": 12} +{"episode_index": 360, "length": 48} +{"episode_index": 361, "length": 44} +{"episode_index": 362, "length": 45} +{"episode_index": 363, "length": 46} +{"episode_index": 364, "length": 18} +{"episode_index": 365, "length": 45} +{"episode_index": 366, "length": 46} +{"episode_index": 367, "length": 34} +{"episode_index": 368, "length": 24} +{"episode_index": 369, "length": 45} +{"episode_index": 370, "length": 44} +{"episode_index": 371, "length": 45} +{"episode_index": 372, "length": 46} +{"episode_index": 373, "length": 46} +{"episode_index": 374, "length": 23} +{"episode_index": 375, "length": 48} +{"episode_index": 376, "length": 45} +{"episode_index": 377, "length": 45} +{"episode_index": 378, "length": 49} +{"episode_index": 379, "length": 43} +{"episode_index": 380, "length": 45} +{"episode_index": 381, "length": 45} +{"episode_index": 382, "length": 13} +{"episode_index": 383, "length": 45} +{"episode_index": 384, "length": 15} +{"episode_index": 385, "length": 45} +{"episode_index": 386, "length": 46} +{"episode_index": 387, "length": 32} +{"episode_index": 388, "length": 45} +{"episode_index": 389, "length": 45} +{"episode_index": 390, "length": 35} +{"episode_index": 391, "length": 44} +{"episode_index": 392, "length": 45} +{"episode_index": 393, "length": 44} +{"episode_index": 394, "length": 45} +{"episode_index": 395, "length": 45} +{"episode_index": 396, "length": 44} +{"episode_index": 397, "length": 45} +{"episode_index": 398, "length": 45} +{"episode_index": 399, "length": 46} +{"episode_index": 400, "length": 45} +{"episode_index": 401, "length": 43} +{"episode_index": 402, "length": 49} +{"episode_index": 403, "length": 44} +{"episode_index": 404, "length": 45} +{"episode_index": 405, "length": 47} +{"episode_index": 406, "length": 43} +{"episode_index": 407, "length": 45} +{"episode_index": 408, "length": 45} +{"episode_index": 409, "length": 45} +{"episode_index": 410, "length": 29} +{"episode_index": 411, "length": 30} +{"episode_index": 412, "length": 44} +{"episode_index": 413, "length": 43} +{"episode_index": 414, "length": 45} +{"episode_index": 415, "length": 45} +{"episode_index": 416, "length": 44} +{"episode_index": 417, "length": 45} +{"episode_index": 418, "length": 45} +{"episode_index": 419, "length": 49} +{"episode_index": 420, "length": 46} +{"episode_index": 421, "length": 43} +{"episode_index": 422, "length": 19} +{"episode_index": 423, "length": 46} +{"episode_index": 424, "length": 48} +{"episode_index": 425, "length": 24} +{"episode_index": 426, "length": 46} +{"episode_index": 427, "length": 45} +{"episode_index": 428, "length": 45} +{"episode_index": 429, "length": 29} +{"episode_index": 430, "length": 45} +{"episode_index": 431, "length": 12} +{"episode_index": 432, "length": 44} +{"episode_index": 433, "length": 44} +{"episode_index": 434, "length": 45} +{"episode_index": 435, "length": 12} +{"episode_index": 436, "length": 33} +{"episode_index": 437, "length": 44} +{"episode_index": 438, "length": 45} +{"episode_index": 439, "length": 45} +{"episode_index": 440, "length": 44} +{"episode_index": 441, "length": 45} +{"episode_index": 442, "length": 45} +{"episode_index": 443, "length": 47} +{"episode_index": 444, "length": 45} +{"episode_index": 445, "length": 46} +{"episode_index": 446, "length": 46} +{"episode_index": 447, "length": 45} +{"episode_index": 448, "length": 47} +{"episode_index": 449, "length": 46} +{"episode_index": 450, "length": 45} +{"episode_index": 451, "length": 47} +{"episode_index": 452, "length": 46} +{"episode_index": 453, "length": 43} +{"episode_index": 454, "length": 45} +{"episode_index": 455, "length": 47} +{"episode_index": 456, "length": 11} +{"episode_index": 457, "length": 13} +{"episode_index": 458, "length": 44} +{"episode_index": 459, "length": 45} +{"episode_index": 460, "length": 45} +{"episode_index": 461, "length": 45} +{"episode_index": 462, "length": 45} +{"episode_index": 463, "length": 45} +{"episode_index": 464, "length": 45} +{"episode_index": 465, "length": 46} +{"episode_index": 466, "length": 12} +{"episode_index": 467, "length": 47} +{"episode_index": 468, "length": 45} +{"episode_index": 469, "length": 44} +{"episode_index": 470, "length": 47} +{"episode_index": 471, "length": 47} +{"episode_index": 472, "length": 28} +{"episode_index": 473, "length": 46} +{"episode_index": 474, "length": 19} +{"episode_index": 475, "length": 43} +{"episode_index": 476, "length": 44} +{"episode_index": 477, "length": 19} +{"episode_index": 478, "length": 13} +{"episode_index": 479, "length": 44} +{"episode_index": 480, "length": 45} +{"episode_index": 481, "length": 44} +{"episode_index": 482, "length": 45} +{"episode_index": 483, "length": 12} +{"episode_index": 484, "length": 45} +{"episode_index": 485, "length": 45} +{"episode_index": 486, "length": 44} +{"episode_index": 487, "length": 45} +{"episode_index": 488, "length": 44} +{"episode_index": 489, "length": 46} +{"episode_index": 490, "length": 45} +{"episode_index": 491, "length": 49} +{"episode_index": 492, "length": 22} +{"episode_index": 493, "length": 49} +{"episode_index": 494, "length": 50} +{"episode_index": 495, "length": 47} +{"episode_index": 496, "length": 46} +{"episode_index": 497, "length": 45} +{"episode_index": 498, "length": 44} +{"episode_index": 499, "length": 46} +{"episode_index": 500, "length": 46} +{"episode_index": 501, "length": 44} +{"episode_index": 502, "length": 46} +{"episode_index": 503, "length": 47} +{"episode_index": 504, "length": 45} +{"episode_index": 505, "length": 48} +{"episode_index": 506, "length": 46} +{"episode_index": 507, "length": 19} +{"episode_index": 508, "length": 45} +{"episode_index": 509, "length": 44} +{"episode_index": 510, "length": 45} +{"episode_index": 511, "length": 24} +{"episode_index": 512, "length": 46} +{"episode_index": 513, "length": 45} +{"episode_index": 514, "length": 45} +{"episode_index": 515, "length": 46} +{"episode_index": 516, "length": 11} +{"episode_index": 517, "length": 45} +{"episode_index": 518, "length": 50} +{"episode_index": 519, "length": 46} +{"episode_index": 520, "length": 45} +{"episode_index": 521, "length": 12} +{"episode_index": 522, "length": 45} +{"episode_index": 523, "length": 44} +{"episode_index": 524, "length": 48} +{"episode_index": 525, "length": 45} +{"episode_index": 526, "length": 45} +{"episode_index": 527, "length": 45} +{"episode_index": 528, "length": 46} +{"episode_index": 529, "length": 44} +{"episode_index": 530, "length": 43} +{"episode_index": 531, "length": 31} +{"episode_index": 532, "length": 45} +{"episode_index": 533, "length": 45} +{"episode_index": 534, "length": 44} +{"episode_index": 535, "length": 44} +{"episode_index": 536, "length": 45} +{"episode_index": 537, "length": 19} +{"episode_index": 538, "length": 43} +{"episode_index": 539, "length": 12} +{"episode_index": 540, "length": 16} +{"episode_index": 541, "length": 19} +{"episode_index": 542, "length": 44} +{"episode_index": 543, "length": 46} +{"episode_index": 544, "length": 45} +{"episode_index": 545, "length": 45} +{"episode_index": 546, "length": 45} +{"episode_index": 547, "length": 45} +{"episode_index": 548, "length": 45} +{"episode_index": 549, "length": 44} +{"episode_index": 550, "length": 45} +{"episode_index": 551, "length": 45} +{"episode_index": 552, "length": 45} +{"episode_index": 553, "length": 43} +{"episode_index": 554, "length": 20} +{"episode_index": 555, "length": 46} +{"episode_index": 556, "length": 45} +{"episode_index": 557, "length": 43} +{"episode_index": 558, "length": 45} +{"episode_index": 559, "length": 4} +{"episode_index": 560, "length": 45} +{"episode_index": 561, "length": 45} +{"episode_index": 562, "length": 44} +{"episode_index": 563, "length": 44} +{"episode_index": 564, "length": 45} +{"episode_index": 565, "length": 47} +{"episode_index": 566, "length": 23} +{"episode_index": 567, "length": 47} +{"episode_index": 568, "length": 47} +{"episode_index": 569, "length": 45} +{"episode_index": 570, "length": 44} +{"episode_index": 571, "length": 46} +{"episode_index": 572, "length": 45} +{"episode_index": 573, "length": 45} +{"episode_index": 574, "length": 47} +{"episode_index": 575, "length": 45} +{"episode_index": 576, "length": 45} +{"episode_index": 577, "length": 44} +{"episode_index": 578, "length": 46} +{"episode_index": 579, "length": 45} +{"episode_index": 580, "length": 45} +{"episode_index": 581, "length": 43} +{"episode_index": 582, "length": 45} +{"episode_index": 583, "length": 48} +{"episode_index": 584, "length": 27} +{"episode_index": 585, "length": 45} +{"episode_index": 586, "length": 45} +{"episode_index": 587, "length": 45} +{"episode_index": 588, "length": 45} +{"episode_index": 589, "length": 22} +{"episode_index": 590, "length": 45} +{"episode_index": 591, "length": 28} +{"episode_index": 592, "length": 44} +{"episode_index": 593, "length": 11} +{"episode_index": 594, "length": 44} +{"episode_index": 595, "length": 46} +{"episode_index": 596, "length": 43} +{"episode_index": 597, "length": 47} +{"episode_index": 598, "length": 45} +{"episode_index": 599, "length": 24} +{"episode_index": 600, "length": 19} +{"episode_index": 601, "length": 45} +{"episode_index": 602, "length": 18} +{"episode_index": 603, "length": 44} +{"episode_index": 604, "length": 46} +{"episode_index": 605, "length": 45} +{"episode_index": 606, "length": 46} +{"episode_index": 607, "length": 44} +{"episode_index": 608, "length": 44} +{"episode_index": 609, "length": 45} +{"episode_index": 610, "length": 48} +{"episode_index": 611, "length": 46} +{"episode_index": 612, "length": 44} +{"episode_index": 613, "length": 45} +{"episode_index": 614, "length": 43} +{"episode_index": 615, "length": 47} +{"episode_index": 616, "length": 46} +{"episode_index": 617, "length": 45} +{"episode_index": 618, "length": 44} +{"episode_index": 619, "length": 45} +{"episode_index": 620, "length": 44} +{"episode_index": 621, "length": 44} +{"episode_index": 622, "length": 45} +{"episode_index": 623, "length": 45} +{"episode_index": 624, "length": 45} +{"episode_index": 625, "length": 19} +{"episode_index": 626, "length": 45} +{"episode_index": 627, "length": 18} +{"episode_index": 628, "length": 12} +{"episode_index": 629, "length": 45} +{"episode_index": 630, "length": 11} +{"episode_index": 631, "length": 45} +{"episode_index": 632, "length": 46} +{"episode_index": 633, "length": 43} +{"episode_index": 634, "length": 43} +{"episode_index": 635, "length": 33} +{"episode_index": 636, "length": 48} +{"episode_index": 637, "length": 47} +{"episode_index": 638, "length": 44} +{"episode_index": 639, "length": 44} +{"episode_index": 640, "length": 45} +{"episode_index": 641, "length": 45} +{"episode_index": 642, "length": 44} +{"episode_index": 643, "length": 45} +{"episode_index": 644, "length": 44} +{"episode_index": 645, "length": 19} +{"episode_index": 646, "length": 45} +{"episode_index": 647, "length": 20} +{"episode_index": 648, "length": 45} +{"episode_index": 649, "length": 46} +{"episode_index": 650, "length": 45} +{"episode_index": 651, "length": 44} +{"episode_index": 652, "length": 44} +{"episode_index": 653, "length": 45} +{"episode_index": 654, "length": 44} +{"episode_index": 655, "length": 45} +{"episode_index": 656, "length": 44} +{"episode_index": 657, "length": 45} +{"episode_index": 658, "length": 45} +{"episode_index": 659, "length": 43} +{"episode_index": 660, "length": 44} +{"episode_index": 661, "length": 44} +{"episode_index": 662, "length": 45} +{"episode_index": 663, "length": 45} +{"episode_index": 664, "length": 46} +{"episode_index": 665, "length": 46} +{"episode_index": 666, "length": 45} +{"episode_index": 667, "length": 45} +{"episode_index": 668, "length": 45} +{"episode_index": 669, "length": 46} +{"episode_index": 670, "length": 44} +{"episode_index": 671, "length": 45} +{"episode_index": 672, "length": 47} +{"episode_index": 673, "length": 43} +{"episode_index": 674, "length": 48} +{"episode_index": 675, "length": 45} +{"episode_index": 676, "length": 11} +{"episode_index": 677, "length": 12} +{"episode_index": 678, "length": 46} +{"episode_index": 679, "length": 44} +{"episode_index": 680, "length": 13} +{"episode_index": 681, "length": 12} +{"episode_index": 682, "length": 19} +{"episode_index": 683, "length": 45} +{"episode_index": 684, "length": 10} +{"episode_index": 685, "length": 45} +{"episode_index": 686, "length": 46} +{"episode_index": 687, "length": 45} +{"episode_index": 688, "length": 45} +{"episode_index": 689, "length": 45} +{"episode_index": 690, "length": 44} +{"episode_index": 691, "length": 45} +{"episode_index": 692, "length": 45} +{"episode_index": 693, "length": 44} +{"episode_index": 694, "length": 46} +{"episode_index": 695, "length": 45} +{"episode_index": 696, "length": 46} +{"episode_index": 697, "length": 44} +{"episode_index": 698, "length": 45} +{"episode_index": 699, "length": 46} +{"episode_index": 700, "length": 45} +{"episode_index": 701, "length": 45} +{"episode_index": 702, "length": 10} +{"episode_index": 703, "length": 33} +{"episode_index": 704, "length": 45} +{"episode_index": 705, "length": 45} +{"episode_index": 706, "length": 45} +{"episode_index": 707, "length": 45} +{"episode_index": 708, "length": 47} +{"episode_index": 709, "length": 47} +{"episode_index": 710, "length": 45} +{"episode_index": 711, "length": 45} +{"episode_index": 712, "length": 47} +{"episode_index": 713, "length": 45} +{"episode_index": 714, "length": 45} +{"episode_index": 715, "length": 19} +{"episode_index": 716, "length": 45} +{"episode_index": 717, "length": 46} +{"episode_index": 718, "length": 45} +{"episode_index": 719, "length": 45} +{"episode_index": 720, "length": 43} +{"episode_index": 721, "length": 44} +{"episode_index": 722, "length": 45} +{"episode_index": 723, "length": 44} +{"episode_index": 724, "length": 45} +{"episode_index": 725, "length": 45} +{"episode_index": 726, "length": 45} +{"episode_index": 727, "length": 45} +{"episode_index": 728, "length": 45} +{"episode_index": 729, "length": 12} +{"episode_index": 730, "length": 44} +{"episode_index": 731, "length": 4} +{"episode_index": 732, "length": 45} +{"episode_index": 733, "length": 46} +{"episode_index": 734, "length": 45} +{"episode_index": 735, "length": 45} +{"episode_index": 736, "length": 45} +{"episode_index": 737, "length": 24} +{"episode_index": 738, "length": 45} +{"episode_index": 739, "length": 44} +{"episode_index": 740, "length": 20} +{"episode_index": 741, "length": 29} +{"episode_index": 742, "length": 45} +{"episode_index": 743, "length": 45} +{"episode_index": 744, "length": 46} +{"episode_index": 745, "length": 46} +{"episode_index": 746, "length": 46} +{"episode_index": 747, "length": 13} +{"episode_index": 748, "length": 46} +{"episode_index": 749, "length": 45} +{"episode_index": 750, "length": 45} +{"episode_index": 751, "length": 44} +{"episode_index": 752, "length": 45} +{"episode_index": 753, "length": 43} +{"episode_index": 754, "length": 45} +{"episode_index": 755, "length": 47} +{"episode_index": 756, "length": 44} +{"episode_index": 757, "length": 45} +{"episode_index": 758, "length": 45} +{"episode_index": 759, "length": 46} +{"episode_index": 760, "length": 44} +{"episode_index": 761, "length": 44} +{"episode_index": 762, "length": 45} +{"episode_index": 763, "length": 43} +{"episode_index": 764, "length": 48} +{"episode_index": 765, "length": 43} +{"episode_index": 766, "length": 44} +{"episode_index": 767, "length": 44} +{"episode_index": 768, "length": 44} +{"episode_index": 769, "length": 44} +{"episode_index": 770, "length": 30} +{"episode_index": 771, "length": 46} +{"episode_index": 772, "length": 45} +{"episode_index": 773, "length": 47} +{"episode_index": 774, "length": 47} +{"episode_index": 775, "length": 43} +{"episode_index": 776, "length": 45} +{"episode_index": 777, "length": 45} +{"episode_index": 778, "length": 43} +{"episode_index": 779, "length": 45} +{"episode_index": 780, "length": 19} +{"episode_index": 781, "length": 46} +{"episode_index": 782, "length": 11} +{"episode_index": 783, "length": 45} +{"episode_index": 784, "length": 44} +{"episode_index": 785, "length": 45} +{"episode_index": 786, "length": 23} +{"episode_index": 787, "length": 45} +{"episode_index": 788, "length": 45} +{"episode_index": 789, "length": 46} +{"episode_index": 790, "length": 43} +{"episode_index": 791, "length": 45} +{"episode_index": 792, "length": 44} +{"episode_index": 793, "length": 12} +{"episode_index": 794, "length": 44} +{"episode_index": 795, "length": 45} +{"episode_index": 796, "length": 32} +{"episode_index": 797, "length": 20} +{"episode_index": 798, "length": 44} +{"episode_index": 799, "length": 43} +{"episode_index": 800, "length": 43} +{"episode_index": 801, "length": 49} +{"episode_index": 802, "length": 45} +{"episode_index": 803, "length": 44} +{"episode_index": 804, "length": 45} +{"episode_index": 805, "length": 47} +{"episode_index": 806, "length": 47} +{"episode_index": 807, "length": 45} +{"episode_index": 808, "length": 45} +{"episode_index": 809, "length": 46} +{"episode_index": 810, "length": 47} +{"episode_index": 811, "length": 46} +{"episode_index": 812, "length": 46} +{"episode_index": 813, "length": 43} +{"episode_index": 814, "length": 44} +{"episode_index": 815, "length": 45} +{"episode_index": 816, "length": 45} +{"episode_index": 817, "length": 44} +{"episode_index": 818, "length": 43} +{"episode_index": 819, "length": 45} +{"episode_index": 820, "length": 45} +{"episode_index": 821, "length": 44} +{"episode_index": 822, "length": 49} +{"episode_index": 823, "length": 46} +{"episode_index": 824, "length": 44} +{"episode_index": 825, "length": 45} +{"episode_index": 826, "length": 45} +{"episode_index": 827, "length": 45} +{"episode_index": 828, "length": 45} +{"episode_index": 829, "length": 46} +{"episode_index": 830, "length": 12} +{"episode_index": 831, "length": 44} +{"episode_index": 832, "length": 24} +{"episode_index": 833, "length": 45} +{"episode_index": 834, "length": 46} +{"episode_index": 835, "length": 45} +{"episode_index": 836, "length": 19} +{"episode_index": 837, "length": 22} +{"episode_index": 838, "length": 45} +{"episode_index": 839, "length": 44} +{"episode_index": 840, "length": 46} +{"episode_index": 841, "length": 45} +{"episode_index": 842, "length": 19} +{"episode_index": 843, "length": 5} +{"episode_index": 844, "length": 25} +{"episode_index": 845, "length": 45} +{"episode_index": 846, "length": 45} +{"episode_index": 847, "length": 4} +{"episode_index": 848, "length": 45} +{"episode_index": 849, "length": 45} +{"episode_index": 850, "length": 45} +{"episode_index": 851, "length": 43} +{"episode_index": 852, "length": 46} +{"episode_index": 853, "length": 46} +{"episode_index": 854, "length": 45} +{"episode_index": 855, "length": 44} +{"episode_index": 856, "length": 45} +{"episode_index": 857, "length": 46} +{"episode_index": 858, "length": 45} +{"episode_index": 859, "length": 46} +{"episode_index": 860, "length": 45} +{"episode_index": 861, "length": 47} +{"episode_index": 862, "length": 45} +{"episode_index": 863, "length": 11} +{"episode_index": 864, "length": 44} +{"episode_index": 865, "length": 31} +{"episode_index": 866, "length": 46} +{"episode_index": 867, "length": 11} +{"episode_index": 868, "length": 16} +{"episode_index": 869, "length": 45} +{"episode_index": 870, "length": 45} +{"episode_index": 871, "length": 45} +{"episode_index": 872, "length": 35} +{"episode_index": 873, "length": 43} +{"episode_index": 874, "length": 45} +{"episode_index": 875, "length": 45} +{"episode_index": 876, "length": 46} +{"episode_index": 877, "length": 45} +{"episode_index": 878, "length": 43} +{"episode_index": 879, "length": 43} +{"episode_index": 880, "length": 46} +{"episode_index": 881, "length": 45} +{"episode_index": 882, "length": 45} +{"episode_index": 883, "length": 45} +{"episode_index": 884, "length": 26} +{"episode_index": 885, "length": 46} +{"episode_index": 886, "length": 44} +{"episode_index": 887, "length": 49} +{"episode_index": 888, "length": 47} +{"episode_index": 889, "length": 48} +{"episode_index": 890, "length": 45} +{"episode_index": 891, "length": 44} +{"episode_index": 892, "length": 44} +{"episode_index": 893, "length": 45} +{"episode_index": 894, "length": 45} +{"episode_index": 895, "length": 44} +{"episode_index": 896, "length": 45} +{"episode_index": 897, "length": 45} +{"episode_index": 898, "length": 45} +{"episode_index": 899, "length": 45} +{"episode_index": 900, "length": 44} +{"episode_index": 901, "length": 43} +{"episode_index": 902, "length": 47} +{"episode_index": 903, "length": 46} +{"episode_index": 904, "length": 45} +{"episode_index": 905, "length": 43} +{"episode_index": 906, "length": 43} +{"episode_index": 907, "length": 46} +{"episode_index": 908, "length": 45} +{"episode_index": 909, "length": 46} +{"episode_index": 910, "length": 45} +{"episode_index": 911, "length": 44} +{"episode_index": 912, "length": 46} +{"episode_index": 913, "length": 47} +{"episode_index": 914, "length": 44} +{"episode_index": 915, "length": 45} +{"episode_index": 916, "length": 47} +{"episode_index": 917, "length": 46} +{"episode_index": 918, "length": 43} +{"episode_index": 919, "length": 45} +{"episode_index": 920, "length": 19} +{"episode_index": 921, "length": 46} +{"episode_index": 922, "length": 44} +{"episode_index": 923, "length": 45} +{"episode_index": 924, "length": 46} +{"episode_index": 925, "length": 46} +{"episode_index": 926, "length": 45} +{"episode_index": 927, "length": 45} +{"episode_index": 928, "length": 44} +{"episode_index": 929, "length": 45} +{"episode_index": 930, "length": 45} +{"episode_index": 931, "length": 46} +{"episode_index": 932, "length": 18} +{"episode_index": 933, "length": 46} +{"episode_index": 934, "length": 49} +{"episode_index": 935, "length": 46} +{"episode_index": 936, "length": 45} +{"episode_index": 937, "length": 43} +{"episode_index": 938, "length": 45} +{"episode_index": 939, "length": 46} +{"episode_index": 940, "length": 45} +{"episode_index": 941, "length": 45} +{"episode_index": 942, "length": 44} +{"episode_index": 943, "length": 32} +{"episode_index": 944, "length": 45} +{"episode_index": 945, "length": 45} +{"episode_index": 946, "length": 46} +{"episode_index": 947, "length": 12} +{"episode_index": 948, "length": 45} +{"episode_index": 949, "length": 44} +{"episode_index": 950, "length": 44} +{"episode_index": 951, "length": 46} +{"episode_index": 952, "length": 44} +{"episode_index": 953, "length": 45} +{"episode_index": 954, "length": 45} +{"episode_index": 955, "length": 19} +{"episode_index": 956, "length": 19} +{"episode_index": 957, "length": 49} +{"episode_index": 958, "length": 47} +{"episode_index": 959, "length": 44} +{"episode_index": 960, "length": 50} +{"episode_index": 961, "length": 45} +{"episode_index": 962, "length": 45} +{"episode_index": 963, "length": 44} +{"episode_index": 964, "length": 23} +{"episode_index": 965, "length": 45} +{"episode_index": 966, "length": 18} +{"episode_index": 967, "length": 44} +{"episode_index": 968, "length": 46} +{"episode_index": 969, "length": 45} +{"episode_index": 970, "length": 47} +{"episode_index": 971, "length": 45} +{"episode_index": 972, "length": 47} +{"episode_index": 973, "length": 46} +{"episode_index": 974, "length": 45} +{"episode_index": 975, "length": 11} +{"episode_index": 976, "length": 31} +{"episode_index": 977, "length": 44} +{"episode_index": 978, "length": 11} +{"episode_index": 979, "length": 43} +{"episode_index": 980, "length": 32} +{"episode_index": 981, "length": 44} +{"episode_index": 982, "length": 13} +{"episode_index": 983, "length": 45} +{"episode_index": 984, "length": 46} +{"episode_index": 985, "length": 46} +{"episode_index": 986, "length": 47} +{"episode_index": 987, "length": 45} +{"episode_index": 988, "length": 45} +{"episode_index": 989, "length": 49} +{"episode_index": 990, "length": 44} +{"episode_index": 991, "length": 45} +{"episode_index": 992, "length": 45} +{"episode_index": 993, "length": 45} +{"episode_index": 994, "length": 46} +{"episode_index": 995, "length": 45} +{"episode_index": 996, "length": 45} +{"episode_index": 997, "length": 44} +{"episode_index": 998, "length": 43} +{"episode_index": 999, "length": 44} +{"episode_index": 1000, "length": 23} +{"episode_index": 1001, "length": 46} +{"episode_index": 1002, "length": 45} +{"episode_index": 1003, "length": 15} +{"episode_index": 1004, "length": 45} +{"episode_index": 1005, "length": 45} +{"episode_index": 1006, "length": 5} +{"episode_index": 1007, "length": 16} +{"episode_index": 1008, "length": 46} +{"episode_index": 1009, "length": 44} +{"episode_index": 1010, "length": 46} +{"episode_index": 1011, "length": 45} +{"episode_index": 1012, "length": 46} +{"episode_index": 1013, "length": 44} +{"episode_index": 1014, "length": 12} +{"episode_index": 1015, "length": 44} +{"episode_index": 1016, "length": 44} +{"episode_index": 1017, "length": 45} +{"episode_index": 1018, "length": 30} +{"episode_index": 1019, "length": 20} +{"episode_index": 1020, "length": 45} +{"episode_index": 1021, "length": 44} +{"episode_index": 1022, "length": 23} +{"episode_index": 1023, "length": 44} +{"episode_index": 1024, "length": 45} +{"episode_index": 1025, "length": 45} +{"episode_index": 1026, "length": 45} +{"episode_index": 1027, "length": 45} +{"episode_index": 1028, "length": 45} +{"episode_index": 1029, "length": 44} +{"episode_index": 1030, "length": 15} +{"episode_index": 1031, "length": 16} +{"episode_index": 1032, "length": 49} +{"episode_index": 1033, "length": 44} +{"episode_index": 1034, "length": 43} +{"episode_index": 1035, "length": 46} +{"episode_index": 1036, "length": 44} +{"episode_index": 1037, "length": 20} +{"episode_index": 1038, "length": 46} +{"episode_index": 1039, "length": 45} +{"episode_index": 1040, "length": 44} +{"episode_index": 1041, "length": 44} +{"episode_index": 1042, "length": 45} +{"episode_index": 1043, "length": 44} +{"episode_index": 1044, "length": 45} +{"episode_index": 1045, "length": 25} +{"episode_index": 1046, "length": 44} +{"episode_index": 1047, "length": 43} +{"episode_index": 1048, "length": 44} +{"episode_index": 1049, "length": 46} +{"episode_index": 1050, "length": 45} +{"episode_index": 1051, "length": 45} +{"episode_index": 1052, "length": 44} +{"episode_index": 1053, "length": 44} +{"episode_index": 1054, "length": 45} +{"episode_index": 1055, "length": 19} +{"episode_index": 1056, "length": 46} +{"episode_index": 1057, "length": 46} +{"episode_index": 1058, "length": 11} +{"episode_index": 1059, "length": 45} +{"episode_index": 1060, "length": 47} +{"episode_index": 1061, "length": 45} +{"episode_index": 1062, "length": 47} +{"episode_index": 1063, "length": 45} +{"episode_index": 1064, "length": 45} +{"episode_index": 1065, "length": 47} +{"episode_index": 1066, "length": 22} +{"episode_index": 1067, "length": 43} +{"episode_index": 1068, "length": 44} +{"episode_index": 1069, "length": 45} +{"episode_index": 1070, "length": 45} +{"episode_index": 1071, "length": 44} +{"episode_index": 1072, "length": 19} +{"episode_index": 1073, "length": 45} +{"episode_index": 1074, "length": 19} +{"episode_index": 1075, "length": 12} +{"episode_index": 1076, "length": 5} +{"episode_index": 1077, "length": 45} +{"episode_index": 1078, "length": 44} +{"episode_index": 1079, "length": 46} +{"episode_index": 1080, "length": 45} +{"episode_index": 1081, "length": 45} +{"episode_index": 1082, "length": 45} +{"episode_index": 1083, "length": 47} +{"episode_index": 1084, "length": 44} +{"episode_index": 1085, "length": 45} +{"episode_index": 1086, "length": 44} +{"episode_index": 1087, "length": 43} +{"episode_index": 1088, "length": 45} +{"episode_index": 1089, "length": 47} +{"episode_index": 1090, "length": 18} +{"episode_index": 1091, "length": 21} +{"episode_index": 1092, "length": 45} +{"episode_index": 1093, "length": 11} +{"episode_index": 1094, "length": 43} +{"episode_index": 1095, "length": 19} +{"episode_index": 1096, "length": 44} +{"episode_index": 1097, "length": 45} +{"episode_index": 1098, "length": 45} +{"episode_index": 1099, "length": 45} +{"episode_index": 1100, "length": 16} +{"episode_index": 1101, "length": 45} +{"episode_index": 1102, "length": 46} +{"episode_index": 1103, "length": 43} +{"episode_index": 1104, "length": 45} +{"episode_index": 1105, "length": 45} +{"episode_index": 1106, "length": 46} +{"episode_index": 1107, "length": 46} +{"episode_index": 1108, "length": 45} +{"episode_index": 1109, "length": 44} +{"episode_index": 1110, "length": 45} +{"episode_index": 1111, "length": 44} +{"episode_index": 1112, "length": 44} +{"episode_index": 1113, "length": 45} +{"episode_index": 1114, "length": 44} +{"episode_index": 1115, "length": 45} +{"episode_index": 1116, "length": 46} +{"episode_index": 1117, "length": 12} +{"episode_index": 1118, "length": 47} +{"episode_index": 1119, "length": 44} +{"episode_index": 1120, "length": 46} +{"episode_index": 1121, "length": 45} +{"episode_index": 1122, "length": 43} +{"episode_index": 1123, "length": 19} +{"episode_index": 1124, "length": 45} +{"episode_index": 1125, "length": 47} +{"episode_index": 1126, "length": 45} +{"episode_index": 1127, "length": 45} +{"episode_index": 1128, "length": 45} +{"episode_index": 1129, "length": 44} +{"episode_index": 1130, "length": 45} +{"episode_index": 1131, "length": 43} +{"episode_index": 1132, "length": 45} +{"episode_index": 1133, "length": 46} +{"episode_index": 1134, "length": 44} +{"episode_index": 1135, "length": 47} +{"episode_index": 1136, "length": 45} +{"episode_index": 1137, "length": 45} +{"episode_index": 1138, "length": 47} +{"episode_index": 1139, "length": 45} +{"episode_index": 1140, "length": 45} +{"episode_index": 1141, "length": 46} +{"episode_index": 1142, "length": 44} +{"episode_index": 1143, "length": 45} +{"episode_index": 1144, "length": 45} +{"episode_index": 1145, "length": 45} +{"episode_index": 1146, "length": 44} +{"episode_index": 1147, "length": 45} +{"episode_index": 1148, "length": 45} +{"episode_index": 1149, "length": 47} +{"episode_index": 1150, "length": 45} +{"episode_index": 1151, "length": 45} +{"episode_index": 1152, "length": 45} +{"episode_index": 1153, "length": 44} +{"episode_index": 1154, "length": 19} +{"episode_index": 1155, "length": 45} +{"episode_index": 1156, "length": 45} +{"episode_index": 1157, "length": 45} +{"episode_index": 1158, "length": 11} +{"episode_index": 1159, "length": 45} +{"episode_index": 1160, "length": 44} +{"episode_index": 1161, "length": 44} +{"episode_index": 1162, "length": 44} +{"episode_index": 1163, "length": 25} +{"episode_index": 1164, "length": 45} +{"episode_index": 1165, "length": 44} +{"episode_index": 1166, "length": 45} +{"episode_index": 1167, "length": 44} +{"episode_index": 1168, "length": 45} +{"episode_index": 1169, "length": 43} +{"episode_index": 1170, "length": 45} +{"episode_index": 1171, "length": 52} +{"episode_index": 1172, "length": 44} +{"episode_index": 1173, "length": 48} +{"episode_index": 1174, "length": 46} +{"episode_index": 1175, "length": 45} +{"episode_index": 1176, "length": 11} +{"episode_index": 1177, "length": 43} +{"episode_index": 1178, "length": 46} +{"episode_index": 1179, "length": 45} +{"episode_index": 1180, "length": 45} +{"episode_index": 1181, "length": 45} +{"episode_index": 1182, "length": 43} +{"episode_index": 1183, "length": 46} +{"episode_index": 1184, "length": 20} +{"episode_index": 1185, "length": 44} +{"episode_index": 1186, "length": 46} +{"episode_index": 1187, "length": 44} +{"episode_index": 1188, "length": 44} +{"episode_index": 1189, "length": 45} +{"episode_index": 1190, "length": 45} +{"episode_index": 1191, "length": 45} +{"episode_index": 1192, "length": 46} +{"episode_index": 1193, "length": 46} +{"episode_index": 1194, "length": 46} +{"episode_index": 1195, "length": 44} +{"episode_index": 1196, "length": 23} +{"episode_index": 1197, "length": 45} +{"episode_index": 1198, "length": 44} +{"episode_index": 1199, "length": 45} +{"episode_index": 1200, "length": 43} +{"episode_index": 1201, "length": 44} +{"episode_index": 1202, "length": 45} +{"episode_index": 1203, "length": 18} +{"episode_index": 1204, "length": 45} +{"episode_index": 1205, "length": 45} +{"episode_index": 1206, "length": 45} +{"episode_index": 1207, "length": 47} +{"episode_index": 1208, "length": 47} +{"episode_index": 1209, "length": 44} +{"episode_index": 1210, "length": 45} +{"episode_index": 1211, "length": 12} +{"episode_index": 1212, "length": 45} +{"episode_index": 1213, "length": 45} +{"episode_index": 1214, "length": 45} +{"episode_index": 1215, "length": 43} +{"episode_index": 1216, "length": 44} +{"episode_index": 1217, "length": 44} +{"episode_index": 1218, "length": 45} +{"episode_index": 1219, "length": 31} +{"episode_index": 1220, "length": 43} +{"episode_index": 1221, "length": 44} +{"episode_index": 1222, "length": 28} +{"episode_index": 1223, "length": 45} +{"episode_index": 1224, "length": 45} +{"episode_index": 1225, "length": 44} +{"episode_index": 1226, "length": 45} +{"episode_index": 1227, "length": 43} +{"episode_index": 1228, "length": 49} +{"episode_index": 1229, "length": 45} +{"episode_index": 1230, "length": 18} +{"episode_index": 1231, "length": 28} +{"episode_index": 1232, "length": 46} +{"episode_index": 1233, "length": 47} +{"episode_index": 1234, "length": 45} +{"episode_index": 1235, "length": 44} +{"episode_index": 1236, "length": 44} +{"episode_index": 1237, "length": 45} +{"episode_index": 1238, "length": 44} +{"episode_index": 1239, "length": 11} +{"episode_index": 1240, "length": 45} +{"episode_index": 1241, "length": 46} +{"episode_index": 1242, "length": 44} +{"episode_index": 1243, "length": 31} +{"episode_index": 1244, "length": 45} +{"episode_index": 1245, "length": 46} +{"episode_index": 1246, "length": 45} +{"episode_index": 1247, "length": 46} +{"episode_index": 1248, "length": 44} +{"episode_index": 1249, "length": 45} +{"episode_index": 1250, "length": 45} +{"episode_index": 1251, "length": 46} +{"episode_index": 1252, "length": 45} +{"episode_index": 1253, "length": 45} +{"episode_index": 1254, "length": 45} +{"episode_index": 1255, "length": 45} +{"episode_index": 1256, "length": 44} +{"episode_index": 1257, "length": 45} +{"episode_index": 1258, "length": 43} +{"episode_index": 1259, "length": 44} +{"episode_index": 1260, "length": 44} +{"episode_index": 1261, "length": 46} +{"episode_index": 1262, "length": 12} +{"episode_index": 1263, "length": 45} +{"episode_index": 1264, "length": 44} +{"episode_index": 1265, "length": 45} +{"episode_index": 1266, "length": 45} +{"episode_index": 1267, "length": 45} +{"episode_index": 1268, "length": 46} +{"episode_index": 1269, "length": 45} +{"episode_index": 1270, "length": 45} +{"episode_index": 1271, "length": 45} +{"episode_index": 1272, "length": 32} +{"episode_index": 1273, "length": 20} +{"episode_index": 1274, "length": 31} +{"episode_index": 1275, "length": 45} +{"episode_index": 1276, "length": 48} +{"episode_index": 1277, "length": 46} +{"episode_index": 1278, "length": 50} +{"episode_index": 1279, "length": 44} +{"episode_index": 1280, "length": 45} +{"episode_index": 1281, "length": 18} +{"episode_index": 1282, "length": 44} +{"episode_index": 1283, "length": 45} +{"episode_index": 1284, "length": 44} +{"episode_index": 1285, "length": 43} +{"episode_index": 1286, "length": 45} +{"episode_index": 1287, "length": 45} +{"episode_index": 1288, "length": 11} +{"episode_index": 1289, "length": 46} +{"episode_index": 1290, "length": 45} +{"episode_index": 1291, "length": 44} +{"episode_index": 1292, "length": 43} +{"episode_index": 1293, "length": 45} +{"episode_index": 1294, "length": 43} +{"episode_index": 1295, "length": 45} +{"episode_index": 1296, "length": 46} +{"episode_index": 1297, "length": 45} +{"episode_index": 1298, "length": 44} +{"episode_index": 1299, "length": 45} +{"episode_index": 1300, "length": 22} +{"episode_index": 1301, "length": 45} +{"episode_index": 1302, "length": 45} +{"episode_index": 1303, "length": 45} +{"episode_index": 1304, "length": 43} +{"episode_index": 1305, "length": 44} +{"episode_index": 1306, "length": 46} +{"episode_index": 1307, "length": 45} +{"episode_index": 1308, "length": 18} +{"episode_index": 1309, "length": 45} +{"episode_index": 1310, "length": 44} +{"episode_index": 1311, "length": 45} +{"episode_index": 1312, "length": 45} +{"episode_index": 1313, "length": 45} +{"episode_index": 1314, "length": 46} +{"episode_index": 1315, "length": 45} +{"episode_index": 1316, "length": 45} +{"episode_index": 1317, "length": 45} +{"episode_index": 1318, "length": 45} +{"episode_index": 1319, "length": 46} +{"episode_index": 1320, "length": 45} +{"episode_index": 1321, "length": 45} +{"episode_index": 1322, "length": 45} +{"episode_index": 1323, "length": 3} +{"episode_index": 1324, "length": 33} +{"episode_index": 1325, "length": 45} +{"episode_index": 1326, "length": 45} +{"episode_index": 1327, "length": 46} +{"episode_index": 1328, "length": 44} +{"episode_index": 1329, "length": 45} +{"episode_index": 1330, "length": 51} +{"episode_index": 1331, "length": 46} +{"episode_index": 1332, "length": 45} +{"episode_index": 1333, "length": 26} +{"episode_index": 1334, "length": 44} +{"episode_index": 1335, "length": 44} +{"episode_index": 1336, "length": 44} +{"episode_index": 1337, "length": 32} +{"episode_index": 1338, "length": 45} +{"episode_index": 1339, "length": 43} +{"episode_index": 1340, "length": 44} +{"episode_index": 1341, "length": 45} +{"episode_index": 1342, "length": 44} +{"episode_index": 1343, "length": 12} +{"episode_index": 1344, "length": 44} +{"episode_index": 1345, "length": 12} +{"episode_index": 1346, "length": 46} +{"episode_index": 1347, "length": 46} +{"episode_index": 1348, "length": 45} +{"episode_index": 1349, "length": 45} +{"episode_index": 1350, "length": 25} +{"episode_index": 1351, "length": 45} +{"episode_index": 1352, "length": 39} +{"episode_index": 1353, "length": 45} +{"episode_index": 1354, "length": 48} +{"episode_index": 1355, "length": 44} +{"episode_index": 1356, "length": 44} +{"episode_index": 1357, "length": 45} +{"episode_index": 1358, "length": 44} +{"episode_index": 1359, "length": 45} +{"episode_index": 1360, "length": 44} +{"episode_index": 1361, "length": 45} +{"episode_index": 1362, "length": 19} +{"episode_index": 1363, "length": 46} +{"episode_index": 1364, "length": 44} +{"episode_index": 1365, "length": 48} +{"episode_index": 1366, "length": 45} +{"episode_index": 1367, "length": 47} +{"episode_index": 1368, "length": 43} +{"episode_index": 1369, "length": 47} +{"episode_index": 1370, "length": 18} +{"episode_index": 1371, "length": 43} +{"episode_index": 1372, "length": 45} +{"episode_index": 1373, "length": 11} +{"episode_index": 1374, "length": 45} +{"episode_index": 1375, "length": 11} +{"episode_index": 1376, "length": 18} +{"episode_index": 1377, "length": 44} +{"episode_index": 1378, "length": 43} +{"episode_index": 1379, "length": 46} +{"episode_index": 1380, "length": 44} +{"episode_index": 1381, "length": 10} +{"episode_index": 1382, "length": 45} +{"episode_index": 1383, "length": 45} +{"episode_index": 1384, "length": 45} +{"episode_index": 1385, "length": 45} +{"episode_index": 1386, "length": 43} +{"episode_index": 1387, "length": 46} +{"episode_index": 1388, "length": 43} +{"episode_index": 1389, "length": 44} +{"episode_index": 1390, "length": 44} +{"episode_index": 1391, "length": 47} +{"episode_index": 1392, "length": 47} +{"episode_index": 1393, "length": 43} +{"episode_index": 1394, "length": 45} +{"episode_index": 1395, "length": 44} +{"episode_index": 1396, "length": 17} +{"episode_index": 1397, "length": 45} +{"episode_index": 1398, "length": 45} +{"episode_index": 1399, "length": 45} +{"episode_index": 1400, "length": 46} +{"episode_index": 1401, "length": 45} +{"episode_index": 1402, "length": 45} +{"episode_index": 1403, "length": 45} +{"episode_index": 1404, "length": 45} +{"episode_index": 1405, "length": 45} +{"episode_index": 1406, "length": 46} +{"episode_index": 1407, "length": 46} +{"episode_index": 1408, "length": 46} +{"episode_index": 1409, "length": 18} +{"episode_index": 1410, "length": 44} +{"episode_index": 1411, "length": 44} +{"episode_index": 1412, "length": 48} +{"episode_index": 1413, "length": 45} +{"episode_index": 1414, "length": 50} +{"episode_index": 1415, "length": 12} +{"episode_index": 1416, "length": 46} +{"episode_index": 1417, "length": 45} +{"episode_index": 1418, "length": 19} +{"episode_index": 1419, "length": 45} +{"episode_index": 1420, "length": 45} +{"episode_index": 1421, "length": 44} +{"episode_index": 1422, "length": 45} +{"episode_index": 1423, "length": 44} +{"episode_index": 1424, "length": 43} +{"episode_index": 1425, "length": 44} +{"episode_index": 1426, "length": 45} +{"episode_index": 1427, "length": 44} +{"episode_index": 1428, "length": 45} +{"episode_index": 1429, "length": 44} +{"episode_index": 1430, "length": 11} +{"episode_index": 1431, "length": 19} +{"episode_index": 1432, "length": 43} +{"episode_index": 1433, "length": 44} +{"episode_index": 1434, "length": 47} +{"episode_index": 1435, "length": 45} +{"episode_index": 1436, "length": 33} +{"episode_index": 1437, "length": 13} +{"episode_index": 1438, "length": 45} +{"episode_index": 1439, "length": 18} +{"episode_index": 1440, "length": 45} +{"episode_index": 1441, "length": 44} +{"episode_index": 1442, "length": 27} +{"episode_index": 1443, "length": 46} +{"episode_index": 1444, "length": 45} +{"episode_index": 1445, "length": 44} +{"episode_index": 1446, "length": 45} +{"episode_index": 1447, "length": 45} +{"episode_index": 1448, "length": 12} +{"episode_index": 1449, "length": 44} +{"episode_index": 1450, "length": 43} +{"episode_index": 1451, "length": 45} +{"episode_index": 1452, "length": 44} +{"episode_index": 1453, "length": 18} +{"episode_index": 1454, "length": 44} +{"episode_index": 1455, "length": 46} +{"episode_index": 1456, "length": 46} +{"episode_index": 1457, "length": 45} +{"episode_index": 1458, "length": 43} +{"episode_index": 1459, "length": 45} +{"episode_index": 1460, "length": 45} +{"episode_index": 1461, "length": 44} +{"episode_index": 1462, "length": 47} +{"episode_index": 1463, "length": 45} +{"episode_index": 1464, "length": 44} +{"episode_index": 1465, "length": 44} +{"episode_index": 1466, "length": 5} +{"episode_index": 1467, "length": 44} +{"episode_index": 1468, "length": 12} +{"episode_index": 1469, "length": 44} +{"episode_index": 1470, "length": 21} +{"episode_index": 1471, "length": 44} +{"episode_index": 1472, "length": 10} +{"episode_index": 1473, "length": 44} +{"episode_index": 1474, "length": 45} +{"episode_index": 1475, "length": 48} +{"episode_index": 1476, "length": 45} +{"episode_index": 1477, "length": 45} +{"episode_index": 1478, "length": 43} +{"episode_index": 1479, "length": 19} +{"episode_index": 1480, "length": 49} +{"episode_index": 1481, "length": 44} +{"episode_index": 1482, "length": 44} +{"episode_index": 1483, "length": 45} +{"episode_index": 1484, "length": 46} +{"episode_index": 1485, "length": 45} +{"episode_index": 1486, "length": 45} +{"episode_index": 1487, "length": 46} +{"episode_index": 1488, "length": 45} +{"episode_index": 1489, "length": 46} +{"episode_index": 1490, "length": 19} +{"episode_index": 1491, "length": 48} +{"episode_index": 1492, "length": 33} +{"episode_index": 1493, "length": 44} +{"episode_index": 1494, "length": 44} +{"episode_index": 1495, "length": 44} +{"episode_index": 1496, "length": 46} +{"episode_index": 1497, "length": 45} +{"episode_index": 1498, "length": 45} +{"episode_index": 1499, "length": 45} +{"episode_index": 1500, "length": 44} +{"episode_index": 1501, "length": 45} +{"episode_index": 1502, "length": 18} +{"episode_index": 1503, "length": 44} +{"episode_index": 1504, "length": 43} +{"episode_index": 1505, "length": 45} +{"episode_index": 1506, "length": 45} +{"episode_index": 1507, "length": 46} +{"episode_index": 1508, "length": 47} +{"episode_index": 1509, "length": 45} +{"episode_index": 1510, "length": 45} +{"episode_index": 1511, "length": 38} +{"episode_index": 1512, "length": 45} +{"episode_index": 1513, "length": 46} +{"episode_index": 1514, "length": 12} +{"episode_index": 1515, "length": 45} +{"episode_index": 1516, "length": 45} +{"episode_index": 1517, "length": 45} +{"episode_index": 1518, "length": 43} +{"episode_index": 1519, "length": 46} +{"episode_index": 1520, "length": 43} +{"episode_index": 1521, "length": 45} +{"episode_index": 1522, "length": 47} +{"episode_index": 1523, "length": 44} +{"episode_index": 1524, "length": 45} +{"episode_index": 1525, "length": 45} +{"episode_index": 1526, "length": 45} +{"episode_index": 1527, "length": 45} +{"episode_index": 1528, "length": 45} +{"episode_index": 1529, "length": 50} +{"episode_index": 1530, "length": 47} +{"episode_index": 1531, "length": 46} +{"episode_index": 1532, "length": 46} +{"episode_index": 1533, "length": 46} +{"episode_index": 1534, "length": 43} +{"episode_index": 1535, "length": 47} +{"episode_index": 1536, "length": 45} +{"episode_index": 1537, "length": 45} +{"episode_index": 1538, "length": 44} +{"episode_index": 1539, "length": 45} +{"episode_index": 1540, "length": 44} +{"episode_index": 1541, "length": 46} +{"episode_index": 1542, "length": 44} +{"episode_index": 1543, "length": 47} +{"episode_index": 1544, "length": 45} +{"episode_index": 1545, "length": 45} +{"episode_index": 1546, "length": 46} +{"episode_index": 1547, "length": 45} +{"episode_index": 1548, "length": 43} +{"episode_index": 1549, "length": 46} +{"episode_index": 1550, "length": 46} +{"episode_index": 1551, "length": 44} +{"episode_index": 1552, "length": 46} +{"episode_index": 1553, "length": 45} +{"episode_index": 1554, "length": 45} +{"episode_index": 1555, "length": 46} +{"episode_index": 1556, "length": 45} +{"episode_index": 1557, "length": 44} +{"episode_index": 1558, "length": 46} +{"episode_index": 1559, "length": 46} +{"episode_index": 1560, "length": 48} +{"episode_index": 1561, "length": 46} +{"episode_index": 1562, "length": 47} +{"episode_index": 1563, "length": 45} +{"episode_index": 1564, "length": 46} +{"episode_index": 1565, "length": 44} +{"episode_index": 1566, "length": 44} +{"episode_index": 1567, "length": 45} +{"episode_index": 1568, "length": 44} +{"episode_index": 1569, "length": 18} +{"episode_index": 1570, "length": 46} +{"episode_index": 1571, "length": 46} +{"episode_index": 1572, "length": 46} +{"episode_index": 1573, "length": 46} +{"episode_index": 1574, "length": 46} +{"episode_index": 1575, "length": 45} +{"episode_index": 1576, "length": 49} +{"episode_index": 1577, "length": 45} +{"episode_index": 1578, "length": 45} +{"episode_index": 1579, "length": 44} +{"episode_index": 1580, "length": 43} +{"episode_index": 1581, "length": 49} +{"episode_index": 1582, "length": 45} +{"episode_index": 1583, "length": 11} +{"episode_index": 1584, "length": 45} +{"episode_index": 1585, "length": 45} +{"episode_index": 1586, "length": 44} +{"episode_index": 1587, "length": 44} +{"episode_index": 1588, "length": 44} +{"episode_index": 1589, "length": 46} +{"episode_index": 1590, "length": 44} +{"episode_index": 1591, "length": 44} +{"episode_index": 1592, "length": 45} +{"episode_index": 1593, "length": 45} +{"episode_index": 1594, "length": 49} +{"episode_index": 1595, "length": 48} +{"episode_index": 1596, "length": 18} +{"episode_index": 1597, "length": 47} +{"episode_index": 1598, "length": 49} +{"episode_index": 1599, "length": 45} +{"episode_index": 1600, "length": 45} +{"episode_index": 1601, "length": 45} +{"episode_index": 1602, "length": 43} +{"episode_index": 1603, "length": 45} +{"episode_index": 1604, "length": 44} +{"episode_index": 1605, "length": 44} +{"episode_index": 1606, "length": 46} +{"episode_index": 1607, "length": 11} +{"episode_index": 1608, "length": 46} +{"episode_index": 1609, "length": 44} +{"episode_index": 1610, "length": 21} +{"episode_index": 1611, "length": 44} +{"episode_index": 1612, "length": 23} +{"episode_index": 1613, "length": 19} +{"episode_index": 1614, "length": 19} +{"episode_index": 1615, "length": 45} +{"episode_index": 1616, "length": 45} +{"episode_index": 1617, "length": 33} +{"episode_index": 1618, "length": 43} +{"episode_index": 1619, "length": 45} +{"episode_index": 1620, "length": 32} +{"episode_index": 1621, "length": 45} +{"episode_index": 1622, "length": 44} +{"episode_index": 1623, "length": 43} +{"episode_index": 1624, "length": 44} +{"episode_index": 1625, "length": 45} +{"episode_index": 1626, "length": 45} +{"episode_index": 1627, "length": 46} +{"episode_index": 1628, "length": 45} +{"episode_index": 1629, "length": 37} +{"episode_index": 1630, "length": 44} +{"episode_index": 1631, "length": 26} +{"episode_index": 1632, "length": 18} +{"episode_index": 1633, "length": 45} +{"episode_index": 1634, "length": 10} +{"episode_index": 1635, "length": 44} +{"episode_index": 1636, "length": 45} +{"episode_index": 1637, "length": 45} +{"episode_index": 1638, "length": 45} +{"episode_index": 1639, "length": 46} +{"episode_index": 1640, "length": 26} +{"episode_index": 1641, "length": 45} +{"episode_index": 1642, "length": 19} +{"episode_index": 1643, "length": 47} +{"episode_index": 1644, "length": 45} +{"episode_index": 1645, "length": 45} +{"episode_index": 1646, "length": 43} +{"episode_index": 1647, "length": 45} +{"episode_index": 1648, "length": 44} +{"episode_index": 1649, "length": 45} +{"episode_index": 1650, "length": 44} +{"episode_index": 1651, "length": 46} +{"episode_index": 1652, "length": 45} +{"episode_index": 1653, "length": 12} +{"episode_index": 1654, "length": 45} +{"episode_index": 1655, "length": 46} +{"episode_index": 1656, "length": 12} +{"episode_index": 1657, "length": 47} +{"episode_index": 1658, "length": 24} +{"episode_index": 1659, "length": 45} +{"episode_index": 1660, "length": 44} +{"episode_index": 1661, "length": 12} +{"episode_index": 1662, "length": 44} +{"episode_index": 1663, "length": 20} +{"episode_index": 1664, "length": 44} +{"episode_index": 1665, "length": 45} +{"episode_index": 1666, "length": 24} +{"episode_index": 1667, "length": 45} +{"episode_index": 1668, "length": 45} +{"episode_index": 1669, "length": 45} +{"episode_index": 1670, "length": 43} +{"episode_index": 1671, "length": 12} +{"episode_index": 1672, "length": 45} +{"episode_index": 1673, "length": 44} +{"episode_index": 1674, "length": 45} +{"episode_index": 1675, "length": 44} +{"episode_index": 1676, "length": 46} +{"episode_index": 1677, "length": 45} +{"episode_index": 1678, "length": 44} +{"episode_index": 1679, "length": 45} +{"episode_index": 1680, "length": 49} +{"episode_index": 1681, "length": 46} +{"episode_index": 1682, "length": 45} +{"episode_index": 1683, "length": 47} +{"episode_index": 1684, "length": 12} +{"episode_index": 1685, "length": 44} +{"episode_index": 1686, "length": 5} +{"episode_index": 1687, "length": 45} +{"episode_index": 1688, "length": 11} +{"episode_index": 1689, "length": 47} +{"episode_index": 1690, "length": 45} +{"episode_index": 1691, "length": 45} +{"episode_index": 1692, "length": 45} +{"episode_index": 1693, "length": 49} +{"episode_index": 1694, "length": 45} +{"episode_index": 1695, "length": 26} +{"episode_index": 1696, "length": 44} +{"episode_index": 1697, "length": 4} +{"episode_index": 1698, "length": 44} +{"episode_index": 1699, "length": 45} +{"episode_index": 1700, "length": 45} +{"episode_index": 1701, "length": 45} +{"episode_index": 1702, "length": 45} +{"episode_index": 1703, "length": 45} +{"episode_index": 1704, "length": 47} +{"episode_index": 1705, "length": 45} +{"episode_index": 1706, "length": 45} +{"episode_index": 1707, "length": 45} +{"episode_index": 1708, "length": 19} +{"episode_index": 1709, "length": 18} +{"episode_index": 1710, "length": 44} +{"episode_index": 1711, "length": 45} +{"episode_index": 1712, "length": 48} +{"episode_index": 1713, "length": 45} +{"episode_index": 1714, "length": 11} +{"episode_index": 1715, "length": 33} +{"episode_index": 1716, "length": 45} +{"episode_index": 1717, "length": 45} +{"episode_index": 1718, "length": 43} +{"episode_index": 1719, "length": 45} +{"episode_index": 1720, "length": 47} +{"episode_index": 1721, "length": 49} +{"episode_index": 1722, "length": 44} +{"episode_index": 1723, "length": 45} +{"episode_index": 1724, "length": 47} +{"episode_index": 1725, "length": 46} +{"episode_index": 1726, "length": 45} +{"episode_index": 1727, "length": 45} +{"episode_index": 1728, "length": 47} +{"episode_index": 1729, "length": 44} +{"episode_index": 1730, "length": 44} +{"episode_index": 1731, "length": 24} +{"episode_index": 1732, "length": 46} +{"episode_index": 1733, "length": 45} +{"episode_index": 1734, "length": 47} +{"episode_index": 1735, "length": 46} +{"episode_index": 1736, "length": 43} +{"episode_index": 1737, "length": 43} +{"episode_index": 1738, "length": 43} +{"episode_index": 1739, "length": 44} +{"episode_index": 1740, "length": 44} +{"episode_index": 1741, "length": 43} +{"episode_index": 1742, "length": 44} +{"episode_index": 1743, "length": 44} +{"episode_index": 1744, "length": 44} +{"episode_index": 1745, "length": 44} +{"episode_index": 1746, "length": 44} +{"episode_index": 1747, "length": 10} +{"episode_index": 1748, "length": 45} +{"episode_index": 1749, "length": 44} +{"episode_index": 1750, "length": 44} +{"episode_index": 1751, "length": 11} +{"episode_index": 1752, "length": 43} +{"episode_index": 1753, "length": 49} +{"episode_index": 1754, "length": 45} +{"episode_index": 1755, "length": 47} +{"episode_index": 1756, "length": 44} +{"episode_index": 1757, "length": 12} +{"episode_index": 1758, "length": 46} +{"episode_index": 1759, "length": 46} +{"episode_index": 1760, "length": 44} +{"episode_index": 1761, "length": 45} +{"episode_index": 1762, "length": 45} +{"episode_index": 1763, "length": 45} +{"episode_index": 1764, "length": 44} +{"episode_index": 1765, "length": 19} +{"episode_index": 1766, "length": 44} +{"episode_index": 1767, "length": 45} +{"episode_index": 1768, "length": 44} +{"episode_index": 1769, "length": 45} +{"episode_index": 1770, "length": 46} +{"episode_index": 1771, "length": 44} +{"episode_index": 1772, "length": 46} +{"episode_index": 1773, "length": 26} +{"episode_index": 1774, "length": 45} +{"episode_index": 1775, "length": 45} +{"episode_index": 1776, "length": 45} +{"episode_index": 1777, "length": 44} +{"episode_index": 1778, "length": 33} +{"episode_index": 1779, "length": 46} +{"episode_index": 1780, "length": 44} +{"episode_index": 1781, "length": 45} +{"episode_index": 1782, "length": 11} +{"episode_index": 1783, "length": 44} +{"episode_index": 1784, "length": 45} +{"episode_index": 1785, "length": 49} +{"episode_index": 1786, "length": 47} +{"episode_index": 1787, "length": 45} +{"episode_index": 1788, "length": 44} +{"episode_index": 1789, "length": 45} +{"episode_index": 1790, "length": 43} +{"episode_index": 1791, "length": 44} +{"episode_index": 1792, "length": 45} +{"episode_index": 1793, "length": 45} +{"episode_index": 1794, "length": 16} +{"episode_index": 1795, "length": 45} +{"episode_index": 1796, "length": 20} +{"episode_index": 1797, "length": 45} +{"episode_index": 1798, "length": 45} +{"episode_index": 1799, "length": 44} +{"episode_index": 1800, "length": 45} +{"episode_index": 1801, "length": 45} +{"episode_index": 1802, "length": 45} +{"episode_index": 1803, "length": 45} +{"episode_index": 1804, "length": 26} +{"episode_index": 1805, "length": 19} +{"episode_index": 1806, "length": 45} +{"episode_index": 1807, "length": 46} +{"episode_index": 1808, "length": 45} +{"episode_index": 1809, "length": 45} +{"episode_index": 1810, "length": 46} +{"episode_index": 1811, "length": 46} +{"episode_index": 1812, "length": 22} +{"episode_index": 1813, "length": 45} +{"episode_index": 1814, "length": 44} +{"episode_index": 1815, "length": 45} +{"episode_index": 1816, "length": 46} +{"episode_index": 1817, "length": 45} +{"episode_index": 1818, "length": 45} +{"episode_index": 1819, "length": 44} +{"episode_index": 1820, "length": 44} +{"episode_index": 1821, "length": 44} +{"episode_index": 1822, "length": 45} +{"episode_index": 1823, "length": 44} +{"episode_index": 1824, "length": 45} +{"episode_index": 1825, "length": 44} +{"episode_index": 1826, "length": 44} +{"episode_index": 1827, "length": 46} +{"episode_index": 1828, "length": 20} +{"episode_index": 1829, "length": 44} +{"episode_index": 1830, "length": 45} +{"episode_index": 1831, "length": 47} +{"episode_index": 1832, "length": 13} +{"episode_index": 1833, "length": 44} +{"episode_index": 1834, "length": 44} +{"episode_index": 1835, "length": 46} +{"episode_index": 1836, "length": 45} +{"episode_index": 1837, "length": 49} +{"episode_index": 1838, "length": 18} +{"episode_index": 1839, "length": 48} +{"episode_index": 1840, "length": 45} +{"episode_index": 1841, "length": 45} +{"episode_index": 1842, "length": 46} +{"episode_index": 1843, "length": 45} +{"episode_index": 1844, "length": 44} +{"episode_index": 1845, "length": 46} +{"episode_index": 1846, "length": 45} +{"episode_index": 1847, "length": 44} +{"episode_index": 1848, "length": 45} +{"episode_index": 1849, "length": 45} +{"episode_index": 1850, "length": 45} +{"episode_index": 1851, "length": 46} +{"episode_index": 1852, "length": 47} +{"episode_index": 1853, "length": 44} +{"episode_index": 1854, "length": 46} +{"episode_index": 1855, "length": 45} +{"episode_index": 1856, "length": 43} +{"episode_index": 1857, "length": 45} +{"episode_index": 1858, "length": 28} +{"episode_index": 1859, "length": 17} +{"episode_index": 1860, "length": 45} +{"episode_index": 1861, "length": 43} +{"episode_index": 1862, "length": 45} +{"episode_index": 1863, "length": 45} +{"episode_index": 1864, "length": 45} +{"episode_index": 1865, "length": 46} +{"episode_index": 1866, "length": 45} +{"episode_index": 1867, "length": 43} +{"episode_index": 1868, "length": 45} +{"episode_index": 1869, "length": 44} +{"episode_index": 1870, "length": 43} +{"episode_index": 1871, "length": 44} +{"episode_index": 1872, "length": 45} +{"episode_index": 1873, "length": 46} +{"episode_index": 1874, "length": 44} +{"episode_index": 1875, "length": 44} +{"episode_index": 1876, "length": 45} +{"episode_index": 1877, "length": 45} +{"episode_index": 1878, "length": 43} +{"episode_index": 1879, "length": 12} +{"episode_index": 1880, "length": 45} +{"episode_index": 1881, "length": 45} +{"episode_index": 1882, "length": 46} +{"episode_index": 1883, "length": 47} +{"episode_index": 1884, "length": 46} +{"episode_index": 1885, "length": 46} +{"episode_index": 1886, "length": 5} +{"episode_index": 1887, "length": 47} +{"episode_index": 1888, "length": 28} +{"episode_index": 1889, "length": 47} +{"episode_index": 1890, "length": 46} +{"episode_index": 1891, "length": 43} +{"episode_index": 1892, "length": 50} +{"episode_index": 1893, "length": 44} +{"episode_index": 1894, "length": 45} +{"episode_index": 1895, "length": 48} +{"episode_index": 1896, "length": 45} +{"episode_index": 1897, "length": 45} +{"episode_index": 1898, "length": 46} +{"episode_index": 1899, "length": 45} +{"episode_index": 1900, "length": 45} +{"episode_index": 1901, "length": 45} +{"episode_index": 1902, "length": 46} +{"episode_index": 1903, "length": 43} +{"episode_index": 1904, "length": 46} +{"episode_index": 1905, "length": 45} +{"episode_index": 1906, "length": 45} +{"episode_index": 1907, "length": 47} +{"episode_index": 1908, "length": 45} +{"episode_index": 1909, "length": 45} +{"episode_index": 1910, "length": 45} +{"episode_index": 1911, "length": 12} +{"episode_index": 1912, "length": 25} +{"episode_index": 1913, "length": 44} +{"episode_index": 1914, "length": 45} +{"episode_index": 1915, "length": 46} +{"episode_index": 1916, "length": 46} +{"episode_index": 1917, "length": 44} +{"episode_index": 1918, "length": 45} +{"episode_index": 1919, "length": 35} +{"episode_index": 1920, "length": 44} +{"episode_index": 1921, "length": 45} +{"episode_index": 1922, "length": 19} +{"episode_index": 1923, "length": 45} +{"episode_index": 1924, "length": 46} +{"episode_index": 1925, "length": 12} +{"episode_index": 1926, "length": 45} +{"episode_index": 1927, "length": 27} +{"episode_index": 1928, "length": 11} +{"episode_index": 1929, "length": 44} +{"episode_index": 1930, "length": 45} +{"episode_index": 1931, "length": 44} +{"episode_index": 1932, "length": 44} +{"episode_index": 1933, "length": 45} +{"episode_index": 1934, "length": 47} +{"episode_index": 1935, "length": 20} +{"episode_index": 1936, "length": 45} +{"episode_index": 1937, "length": 44} +{"episode_index": 1938, "length": 45} +{"episode_index": 1939, "length": 44} +{"episode_index": 1940, "length": 43} +{"episode_index": 1941, "length": 45} +{"episode_index": 1942, "length": 43} +{"episode_index": 1943, "length": 44} +{"episode_index": 1944, "length": 20} +{"episode_index": 1945, "length": 46} +{"episode_index": 1946, "length": 43} +{"episode_index": 1947, "length": 20} +{"episode_index": 1948, "length": 46} +{"episode_index": 1949, "length": 45} +{"episode_index": 1950, "length": 44} +{"episode_index": 1951, "length": 43} +{"episode_index": 1952, "length": 45} +{"episode_index": 1953, "length": 36} +{"episode_index": 1954, "length": 45} +{"episode_index": 1955, "length": 45} +{"episode_index": 1956, "length": 45} +{"episode_index": 1957, "length": 45} +{"episode_index": 1958, "length": 25} +{"episode_index": 1959, "length": 47} +{"episode_index": 1960, "length": 27} +{"episode_index": 1961, "length": 23} +{"episode_index": 1962, "length": 43} +{"episode_index": 1963, "length": 45} +{"episode_index": 1964, "length": 45} +{"episode_index": 1965, "length": 44} +{"episode_index": 1966, "length": 45} +{"episode_index": 1967, "length": 44} +{"episode_index": 1968, "length": 43} +{"episode_index": 1969, "length": 44} +{"episode_index": 1970, "length": 34} +{"episode_index": 1971, "length": 18} +{"episode_index": 1972, "length": 18} +{"episode_index": 1973, "length": 46} +{"episode_index": 1974, "length": 44} +{"episode_index": 1975, "length": 45} +{"episode_index": 1976, "length": 44} +{"episode_index": 1977, "length": 45} +{"episode_index": 1978, "length": 45} +{"episode_index": 1979, "length": 44} +{"episode_index": 1980, "length": 45} +{"episode_index": 1981, "length": 43} +{"episode_index": 1982, "length": 45} +{"episode_index": 1983, "length": 46} +{"episode_index": 1984, "length": 44} +{"episode_index": 1985, "length": 47} +{"episode_index": 1986, "length": 45} +{"episode_index": 1987, "length": 45} +{"episode_index": 1988, "length": 43} +{"episode_index": 1989, "length": 44} +{"episode_index": 1990, "length": 20} +{"episode_index": 1991, "length": 27} +{"episode_index": 1992, "length": 29} +{"episode_index": 1993, "length": 45} +{"episode_index": 1994, "length": 44} +{"episode_index": 1995, "length": 45} +{"episode_index": 1996, "length": 45} +{"episode_index": 1997, "length": 44} +{"episode_index": 1998, "length": 45} +{"episode_index": 1999, "length": 44} +{"episode_index": 2000, "length": 12} +{"episode_index": 2001, "length": 46} +{"episode_index": 2002, "length": 45} +{"episode_index": 2003, "length": 45} +{"episode_index": 2004, "length": 45} +{"episode_index": 2005, "length": 46} +{"episode_index": 2006, "length": 44} +{"episode_index": 2007, "length": 45} +{"episode_index": 2008, "length": 44} +{"episode_index": 2009, "length": 45} +{"episode_index": 2010, "length": 44} +{"episode_index": 2011, "length": 44} +{"episode_index": 2012, "length": 44} +{"episode_index": 2013, "length": 44} +{"episode_index": 2014, "length": 45} +{"episode_index": 2015, "length": 44} +{"episode_index": 2016, "length": 45} +{"episode_index": 2017, "length": 47} +{"episode_index": 2018, "length": 45} +{"episode_index": 2019, "length": 44} +{"episode_index": 2020, "length": 45} +{"episode_index": 2021, "length": 46} +{"episode_index": 2022, "length": 45} +{"episode_index": 2023, "length": 45} +{"episode_index": 2024, "length": 44} +{"episode_index": 2025, "length": 44} +{"episode_index": 2026, "length": 44} +{"episode_index": 2027, "length": 43} +{"episode_index": 2028, "length": 18} +{"episode_index": 2029, "length": 45} +{"episode_index": 2030, "length": 45} +{"episode_index": 2031, "length": 45} +{"episode_index": 2032, "length": 12} +{"episode_index": 2033, "length": 45} +{"episode_index": 2034, "length": 46} +{"episode_index": 2035, "length": 45} +{"episode_index": 2036, "length": 45} +{"episode_index": 2037, "length": 45} +{"episode_index": 2038, "length": 47} +{"episode_index": 2039, "length": 49} +{"episode_index": 2040, "length": 44} +{"episode_index": 2041, "length": 45} +{"episode_index": 2042, "length": 47} +{"episode_index": 2043, "length": 19} +{"episode_index": 2044, "length": 46} +{"episode_index": 2045, "length": 45} +{"episode_index": 2046, "length": 45} +{"episode_index": 2047, "length": 45} +{"episode_index": 2048, "length": 47} +{"episode_index": 2049, "length": 43} +{"episode_index": 2050, "length": 46} +{"episode_index": 2051, "length": 18} +{"episode_index": 2052, "length": 46} +{"episode_index": 2053, "length": 44} +{"episode_index": 2054, "length": 45} +{"episode_index": 2055, "length": 45} +{"episode_index": 2056, "length": 47} +{"episode_index": 2057, "length": 48} +{"episode_index": 2058, "length": 46} +{"episode_index": 2059, "length": 45} +{"episode_index": 2060, "length": 45} +{"episode_index": 2061, "length": 44} +{"episode_index": 2062, "length": 44} +{"episode_index": 2063, "length": 45} +{"episode_index": 2064, "length": 47} +{"episode_index": 2065, "length": 45} +{"episode_index": 2066, "length": 45} +{"episode_index": 2067, "length": 47} +{"episode_index": 2068, "length": 45} +{"episode_index": 2069, "length": 43} +{"episode_index": 2070, "length": 19} +{"episode_index": 2071, "length": 44} +{"episode_index": 2072, "length": 45} +{"episode_index": 2073, "length": 45} +{"episode_index": 2074, "length": 45} +{"episode_index": 2075, "length": 43} +{"episode_index": 2076, "length": 45} +{"episode_index": 2077, "length": 45} +{"episode_index": 2078, "length": 43} +{"episode_index": 2079, "length": 46} +{"episode_index": 2080, "length": 45} +{"episode_index": 2081, "length": 11} +{"episode_index": 2082, "length": 44} +{"episode_index": 2083, "length": 47} +{"episode_index": 2084, "length": 45} +{"episode_index": 2085, "length": 43} +{"episode_index": 2086, "length": 44} +{"episode_index": 2087, "length": 44} +{"episode_index": 2088, "length": 43} +{"episode_index": 2089, "length": 44} +{"episode_index": 2090, "length": 43} +{"episode_index": 2091, "length": 44} +{"episode_index": 2092, "length": 36} +{"episode_index": 2093, "length": 45} +{"episode_index": 2094, "length": 47} +{"episode_index": 2095, "length": 45} +{"episode_index": 2096, "length": 20} +{"episode_index": 2097, "length": 45} +{"episode_index": 2098, "length": 44} +{"episode_index": 2099, "length": 45} +{"episode_index": 2100, "length": 45} +{"episode_index": 2101, "length": 44} +{"episode_index": 2102, "length": 45} +{"episode_index": 2103, "length": 46} +{"episode_index": 2104, "length": 44} +{"episode_index": 2105, "length": 46} +{"episode_index": 2106, "length": 45} +{"episode_index": 2107, "length": 43} +{"episode_index": 2108, "length": 45} +{"episode_index": 2109, "length": 30} +{"episode_index": 2110, "length": 45} +{"episode_index": 2111, "length": 44} +{"episode_index": 2112, "length": 45} +{"episode_index": 2113, "length": 23} +{"episode_index": 2114, "length": 48} +{"episode_index": 2115, "length": 45} +{"episode_index": 2116, "length": 45} +{"episode_index": 2117, "length": 45} +{"episode_index": 2118, "length": 45} +{"episode_index": 2119, "length": 47} +{"episode_index": 2120, "length": 45} +{"episode_index": 2121, "length": 45} +{"episode_index": 2122, "length": 45} +{"episode_index": 2123, "length": 19} +{"episode_index": 2124, "length": 4} +{"episode_index": 2125, "length": 32} +{"episode_index": 2126, "length": 46} +{"episode_index": 2127, "length": 49} +{"episode_index": 2128, "length": 43} +{"episode_index": 2129, "length": 45} +{"episode_index": 2130, "length": 44} +{"episode_index": 2131, "length": 31} +{"episode_index": 2132, "length": 46} +{"episode_index": 2133, "length": 11} +{"episode_index": 2134, "length": 45} +{"episode_index": 2135, "length": 47} +{"episode_index": 2136, "length": 45} +{"episode_index": 2137, "length": 50} +{"episode_index": 2138, "length": 46} +{"episode_index": 2139, "length": 45} +{"episode_index": 2140, "length": 45} +{"episode_index": 2141, "length": 45} +{"episode_index": 2142, "length": 48} +{"episode_index": 2143, "length": 44} +{"episode_index": 2144, "length": 45} +{"episode_index": 2145, "length": 45} +{"episode_index": 2146, "length": 45} +{"episode_index": 2147, "length": 45} +{"episode_index": 2148, "length": 45} +{"episode_index": 2149, "length": 45} +{"episode_index": 2150, "length": 20} +{"episode_index": 2151, "length": 45} +{"episode_index": 2152, "length": 44} +{"episode_index": 2153, "length": 44} +{"episode_index": 2154, "length": 44} +{"episode_index": 2155, "length": 44} +{"episode_index": 2156, "length": 44} +{"episode_index": 2157, "length": 45} +{"episode_index": 2158, "length": 45} +{"episode_index": 2159, "length": 44} +{"episode_index": 2160, "length": 45} +{"episode_index": 2161, "length": 46} +{"episode_index": 2162, "length": 10} +{"episode_index": 2163, "length": 45} +{"episode_index": 2164, "length": 45} +{"episode_index": 2165, "length": 13} +{"episode_index": 2166, "length": 45} +{"episode_index": 2167, "length": 18} +{"episode_index": 2168, "length": 45} +{"episode_index": 2169, "length": 43} +{"episode_index": 2170, "length": 45} +{"episode_index": 2171, "length": 45} +{"episode_index": 2172, "length": 45} +{"episode_index": 2173, "length": 44} +{"episode_index": 2174, "length": 44} +{"episode_index": 2175, "length": 44} +{"episode_index": 2176, "length": 45} +{"episode_index": 2177, "length": 45} +{"episode_index": 2178, "length": 45} +{"episode_index": 2179, "length": 44} +{"episode_index": 2180, "length": 44} +{"episode_index": 2181, "length": 21} +{"episode_index": 2182, "length": 32} +{"episode_index": 2183, "length": 45} +{"episode_index": 2184, "length": 12} +{"episode_index": 2185, "length": 45} +{"episode_index": 2186, "length": 33} +{"episode_index": 2187, "length": 48} +{"episode_index": 2188, "length": 43} +{"episode_index": 2189, "length": 45} +{"episode_index": 2190, "length": 45} +{"episode_index": 2191, "length": 45} +{"episode_index": 2192, "length": 16} +{"episode_index": 2193, "length": 11} +{"episode_index": 2194, "length": 45} +{"episode_index": 2195, "length": 10} +{"episode_index": 2196, "length": 44} +{"episode_index": 2197, "length": 47} +{"episode_index": 2198, "length": 44} +{"episode_index": 2199, "length": 46} +{"episode_index": 2200, "length": 45} +{"episode_index": 2201, "length": 46} +{"episode_index": 2202, "length": 28} +{"episode_index": 2203, "length": 44} +{"episode_index": 2204, "length": 44} +{"episode_index": 2205, "length": 45} +{"episode_index": 2206, "length": 44} +{"episode_index": 2207, "length": 46} +{"episode_index": 2208, "length": 47} +{"episode_index": 2209, "length": 44} +{"episode_index": 2210, "length": 47} +{"episode_index": 2211, "length": 44} +{"episode_index": 2212, "length": 45} +{"episode_index": 2213, "length": 45} +{"episode_index": 2214, "length": 44} +{"episode_index": 2215, "length": 45} +{"episode_index": 2216, "length": 45} +{"episode_index": 2217, "length": 44} +{"episode_index": 2218, "length": 46} +{"episode_index": 2219, "length": 46} +{"episode_index": 2220, "length": 22} +{"episode_index": 2221, "length": 46} +{"episode_index": 2222, "length": 46} +{"episode_index": 2223, "length": 45} +{"episode_index": 2224, "length": 45} +{"episode_index": 2225, "length": 18} +{"episode_index": 2226, "length": 47} +{"episode_index": 2227, "length": 46} +{"episode_index": 2228, "length": 45} +{"episode_index": 2229, "length": 47} +{"episode_index": 2230, "length": 44} +{"episode_index": 2231, "length": 11} +{"episode_index": 2232, "length": 45} +{"episode_index": 2233, "length": 46} +{"episode_index": 2234, "length": 46} +{"episode_index": 2235, "length": 45} +{"episode_index": 2236, "length": 44} +{"episode_index": 2237, "length": 20} +{"episode_index": 2238, "length": 19} +{"episode_index": 2239, "length": 43} +{"episode_index": 2240, "length": 46} +{"episode_index": 2241, "length": 46} +{"episode_index": 2242, "length": 45} +{"episode_index": 2243, "length": 45} +{"episode_index": 2244, "length": 45} +{"episode_index": 2245, "length": 50} +{"episode_index": 2246, "length": 45} +{"episode_index": 2247, "length": 46} +{"episode_index": 2248, "length": 45} +{"episode_index": 2249, "length": 44} +{"episode_index": 2250, "length": 45} +{"episode_index": 2251, "length": 43} +{"episode_index": 2252, "length": 45} +{"episode_index": 2253, "length": 45} +{"episode_index": 2254, "length": 11} +{"episode_index": 2255, "length": 45} +{"episode_index": 2256, "length": 44} +{"episode_index": 2257, "length": 44} +{"episode_index": 2258, "length": 19} +{"episode_index": 2259, "length": 44} +{"episode_index": 2260, "length": 45} +{"episode_index": 2261, "length": 45} +{"episode_index": 2262, "length": 44} +{"episode_index": 2263, "length": 45} +{"episode_index": 2264, "length": 45} +{"episode_index": 2265, "length": 46} +{"episode_index": 2266, "length": 45} +{"episode_index": 2267, "length": 45} +{"episode_index": 2268, "length": 44} +{"episode_index": 2269, "length": 44} +{"episode_index": 2270, "length": 22} +{"episode_index": 2271, "length": 45} +{"episode_index": 2272, "length": 48} +{"episode_index": 2273, "length": 45} +{"episode_index": 2274, "length": 45} +{"episode_index": 2275, "length": 44} +{"episode_index": 2276, "length": 47} +{"episode_index": 2277, "length": 45} +{"episode_index": 2278, "length": 44} +{"episode_index": 2279, "length": 45} +{"episode_index": 2280, "length": 45} +{"episode_index": 2281, "length": 44} +{"episode_index": 2282, "length": 45} +{"episode_index": 2283, "length": 45} +{"episode_index": 2284, "length": 45} +{"episode_index": 2285, "length": 19} +{"episode_index": 2286, "length": 46} +{"episode_index": 2287, "length": 46} +{"episode_index": 2288, "length": 45} +{"episode_index": 2289, "length": 45} +{"episode_index": 2290, "length": 45} +{"episode_index": 2291, "length": 45} +{"episode_index": 2292, "length": 45} +{"episode_index": 2293, "length": 43} +{"episode_index": 2294, "length": 11} +{"episode_index": 2295, "length": 45} +{"episode_index": 2296, "length": 46} +{"episode_index": 2297, "length": 45} +{"episode_index": 2298, "length": 45} +{"episode_index": 2299, "length": 45} +{"episode_index": 2300, "length": 45} +{"episode_index": 2301, "length": 45} +{"episode_index": 2302, "length": 19} +{"episode_index": 2303, "length": 42} +{"episode_index": 2304, "length": 45} +{"episode_index": 2305, "length": 45} +{"episode_index": 2306, "length": 12} +{"episode_index": 2307, "length": 19} +{"episode_index": 2308, "length": 45} +{"episode_index": 2309, "length": 45} +{"episode_index": 2310, "length": 45} +{"episode_index": 2311, "length": 45} +{"episode_index": 2312, "length": 44} +{"episode_index": 2313, "length": 44} +{"episode_index": 2314, "length": 45} +{"episode_index": 2315, "length": 44} +{"episode_index": 2316, "length": 32} +{"episode_index": 2317, "length": 45} +{"episode_index": 2318, "length": 45} +{"episode_index": 2319, "length": 18} +{"episode_index": 2320, "length": 45} +{"episode_index": 2321, "length": 11} +{"episode_index": 2322, "length": 44} +{"episode_index": 2323, "length": 45} +{"episode_index": 2324, "length": 45} +{"episode_index": 2325, "length": 20} +{"episode_index": 2326, "length": 46} +{"episode_index": 2327, "length": 13} +{"episode_index": 2328, "length": 45} +{"episode_index": 2329, "length": 44} +{"episode_index": 2330, "length": 45} +{"episode_index": 2331, "length": 45} +{"episode_index": 2332, "length": 46} +{"episode_index": 2333, "length": 45} +{"episode_index": 2334, "length": 44} +{"episode_index": 2335, "length": 44} +{"episode_index": 2336, "length": 46} +{"episode_index": 2337, "length": 46} +{"episode_index": 2338, "length": 44} +{"episode_index": 2339, "length": 46} +{"episode_index": 2340, "length": 45} +{"episode_index": 2341, "length": 45} +{"episode_index": 2342, "length": 47} +{"episode_index": 2343, "length": 44} +{"episode_index": 2344, "length": 45} +{"episode_index": 2345, "length": 45} +{"episode_index": 2346, "length": 43} +{"episode_index": 2347, "length": 34} +{"episode_index": 2348, "length": 45} +{"episode_index": 2349, "length": 45} +{"episode_index": 2350, "length": 45} +{"episode_index": 2351, "length": 44} +{"episode_index": 2352, "length": 31} +{"episode_index": 2353, "length": 45} +{"episode_index": 2354, "length": 48} +{"episode_index": 2355, "length": 45} +{"episode_index": 2356, "length": 26} +{"episode_index": 2357, "length": 46} +{"episode_index": 2358, "length": 45} +{"episode_index": 2359, "length": 45} +{"episode_index": 2360, "length": 44} +{"episode_index": 2361, "length": 45} +{"episode_index": 2362, "length": 18} +{"episode_index": 2363, "length": 45} +{"episode_index": 2364, "length": 43} +{"episode_index": 2365, "length": 45} +{"episode_index": 2366, "length": 44} +{"episode_index": 2367, "length": 11} +{"episode_index": 2368, "length": 45} +{"episode_index": 2369, "length": 46} +{"episode_index": 2370, "length": 49} +{"episode_index": 2371, "length": 43} +{"episode_index": 2372, "length": 25} +{"episode_index": 2373, "length": 45} +{"episode_index": 2374, "length": 45} +{"episode_index": 2375, "length": 43} +{"episode_index": 2376, "length": 45} +{"episode_index": 2377, "length": 44} +{"episode_index": 2378, "length": 45} +{"episode_index": 2379, "length": 44} +{"episode_index": 2380, "length": 45} +{"episode_index": 2381, "length": 12} +{"episode_index": 2382, "length": 45} +{"episode_index": 2383, "length": 45} +{"episode_index": 2384, "length": 43} +{"episode_index": 2385, "length": 45} +{"episode_index": 2386, "length": 32} +{"episode_index": 2387, "length": 44} +{"episode_index": 2388, "length": 44} +{"episode_index": 2389, "length": 47} +{"episode_index": 2390, "length": 11} +{"episode_index": 2391, "length": 45} +{"episode_index": 2392, "length": 44} +{"episode_index": 2393, "length": 45} +{"episode_index": 2394, "length": 44} +{"episode_index": 2395, "length": 45} +{"episode_index": 2396, "length": 45} +{"episode_index": 2397, "length": 44} +{"episode_index": 2398, "length": 45} +{"episode_index": 2399, "length": 44} +{"episode_index": 2400, "length": 45} +{"episode_index": 2401, "length": 45} +{"episode_index": 2402, "length": 45} +{"episode_index": 2403, "length": 47} +{"episode_index": 2404, "length": 45} +{"episode_index": 2405, "length": 11} +{"episode_index": 2406, "length": 44} +{"episode_index": 2407, "length": 45} +{"episode_index": 2408, "length": 28} +{"episode_index": 2409, "length": 20} +{"episode_index": 2410, "length": 45} +{"episode_index": 2411, "length": 45} +{"episode_index": 2412, "length": 44} +{"episode_index": 2413, "length": 45} +{"episode_index": 2414, "length": 30} +{"episode_index": 2415, "length": 47} +{"episode_index": 2416, "length": 26} +{"episode_index": 2417, "length": 20} +{"episode_index": 2418, "length": 45} +{"episode_index": 2419, "length": 23} +{"episode_index": 2420, "length": 45} +{"episode_index": 2421, "length": 15} +{"episode_index": 2422, "length": 31} +{"episode_index": 2423, "length": 44} +{"episode_index": 2424, "length": 44} +{"episode_index": 2425, "length": 44} +{"episode_index": 2426, "length": 45} +{"episode_index": 2427, "length": 44} +{"episode_index": 2428, "length": 12} +{"episode_index": 2429, "length": 45} +{"episode_index": 2430, "length": 45} +{"episode_index": 2431, "length": 45} +{"episode_index": 2432, "length": 45} +{"episode_index": 2433, "length": 46} +{"episode_index": 2434, "length": 45} +{"episode_index": 2435, "length": 47} +{"episode_index": 2436, "length": 46} +{"episode_index": 2437, "length": 45} +{"episode_index": 2438, "length": 45} +{"episode_index": 2439, "length": 45} +{"episode_index": 2440, "length": 45} +{"episode_index": 2441, "length": 44} +{"episode_index": 2442, "length": 43} +{"episode_index": 2443, "length": 46} +{"episode_index": 2444, "length": 45} +{"episode_index": 2445, "length": 45} +{"episode_index": 2446, "length": 45} +{"episode_index": 2447, "length": 19} +{"episode_index": 2448, "length": 47} +{"episode_index": 2449, "length": 45} +{"episode_index": 2450, "length": 44} +{"episode_index": 2451, "length": 46} +{"episode_index": 2452, "length": 46} +{"episode_index": 2453, "length": 43} +{"episode_index": 2454, "length": 45} +{"episode_index": 2455, "length": 18} +{"episode_index": 2456, "length": 46} +{"episode_index": 2457, "length": 45} +{"episode_index": 2458, "length": 45} +{"episode_index": 2459, "length": 43} +{"episode_index": 2460, "length": 47} +{"episode_index": 2461, "length": 44} +{"episode_index": 2462, "length": 47} +{"episode_index": 2463, "length": 46} +{"episode_index": 2464, "length": 19} +{"episode_index": 2465, "length": 19} +{"episode_index": 2466, "length": 44} +{"episode_index": 2467, "length": 45} +{"episode_index": 2468, "length": 45} +{"episode_index": 2469, "length": 29} +{"episode_index": 2470, "length": 10} +{"episode_index": 2471, "length": 45} +{"episode_index": 2472, "length": 45} +{"episode_index": 2473, "length": 44} +{"episode_index": 2474, "length": 36} +{"episode_index": 2475, "length": 46} +{"episode_index": 2476, "length": 44} +{"episode_index": 2477, "length": 46} +{"episode_index": 2478, "length": 45} +{"episode_index": 2479, "length": 45} +{"episode_index": 2480, "length": 44} +{"episode_index": 2481, "length": 44} +{"episode_index": 2482, "length": 45} +{"episode_index": 2483, "length": 45} +{"episode_index": 2484, "length": 45} +{"episode_index": 2485, "length": 45} +{"episode_index": 2486, "length": 46} +{"episode_index": 2487, "length": 44} +{"episode_index": 2488, "length": 12} +{"episode_index": 2489, "length": 46} +{"episode_index": 2490, "length": 48} +{"episode_index": 2491, "length": 44} +{"episode_index": 2492, "length": 45} +{"episode_index": 2493, "length": 46} +{"episode_index": 2494, "length": 45} +{"episode_index": 2495, "length": 46} +{"episode_index": 2496, "length": 33} +{"episode_index": 2497, "length": 18} +{"episode_index": 2498, "length": 44} +{"episode_index": 2499, "length": 46} +{"episode_index": 2500, "length": 45} +{"episode_index": 2501, "length": 13} +{"episode_index": 2502, "length": 45} +{"episode_index": 2503, "length": 12} +{"episode_index": 2504, "length": 45} +{"episode_index": 2505, "length": 45} +{"episode_index": 2506, "length": 46} +{"episode_index": 2507, "length": 47} +{"episode_index": 2508, "length": 47} +{"episode_index": 2509, "length": 46} +{"episode_index": 2510, "length": 20} +{"episode_index": 2511, "length": 44} +{"episode_index": 2512, "length": 44} +{"episode_index": 2513, "length": 46} +{"episode_index": 2514, "length": 47} +{"episode_index": 2515, "length": 45} +{"episode_index": 2516, "length": 44} +{"episode_index": 2517, "length": 44} +{"episode_index": 2518, "length": 45} +{"episode_index": 2519, "length": 45} +{"episode_index": 2520, "length": 19} +{"episode_index": 2521, "length": 45} +{"episode_index": 2522, "length": 47} +{"episode_index": 2523, "length": 45} +{"episode_index": 2524, "length": 45} +{"episode_index": 2525, "length": 44} +{"episode_index": 2526, "length": 48} +{"episode_index": 2527, "length": 45} +{"episode_index": 2528, "length": 20} +{"episode_index": 2529, "length": 46} +{"episode_index": 2530, "length": 45} +{"episode_index": 2531, "length": 45} +{"episode_index": 2532, "length": 46} +{"episode_index": 2533, "length": 19} +{"episode_index": 2534, "length": 44} +{"episode_index": 2535, "length": 18} +{"episode_index": 2536, "length": 45} +{"episode_index": 2537, "length": 45} +{"episode_index": 2538, "length": 11} +{"episode_index": 2539, "length": 44} +{"episode_index": 2540, "length": 46} +{"episode_index": 2541, "length": 11} +{"episode_index": 2542, "length": 45} +{"episode_index": 2543, "length": 45} +{"episode_index": 2544, "length": 47} +{"episode_index": 2545, "length": 45} +{"episode_index": 2546, "length": 45} +{"episode_index": 2547, "length": 45} +{"episode_index": 2548, "length": 46} +{"episode_index": 2549, "length": 45} +{"episode_index": 2550, "length": 45} +{"episode_index": 2551, "length": 43} +{"episode_index": 2552, "length": 44} +{"episode_index": 2553, "length": 44} +{"episode_index": 2554, "length": 44} +{"episode_index": 2555, "length": 45} +{"episode_index": 2556, "length": 47} +{"episode_index": 2557, "length": 18} +{"episode_index": 2558, "length": 45} +{"episode_index": 2559, "length": 44} +{"episode_index": 2560, "length": 47} +{"episode_index": 2561, "length": 45} +{"episode_index": 2562, "length": 44} +{"episode_index": 2563, "length": 45} +{"episode_index": 2564, "length": 30} +{"episode_index": 2565, "length": 44} +{"episode_index": 2566, "length": 39} +{"episode_index": 2567, "length": 47} +{"episode_index": 2568, "length": 45} +{"episode_index": 2569, "length": 44} +{"episode_index": 2570, "length": 49} +{"episode_index": 2571, "length": 45} +{"episode_index": 2572, "length": 45} +{"episode_index": 2573, "length": 45} +{"episode_index": 2574, "length": 45} +{"episode_index": 2575, "length": 45} +{"episode_index": 2576, "length": 4} +{"episode_index": 2577, "length": 47} +{"episode_index": 2578, "length": 44} +{"episode_index": 2579, "length": 45} +{"episode_index": 2580, "length": 46} +{"episode_index": 2581, "length": 44} +{"episode_index": 2582, "length": 46} +{"episode_index": 2583, "length": 45} +{"episode_index": 2584, "length": 19} +{"episode_index": 2585, "length": 27} +{"episode_index": 2586, "length": 43} +{"episode_index": 2587, "length": 43} +{"episode_index": 2588, "length": 13} +{"episode_index": 2589, "length": 12} +{"episode_index": 2590, "length": 43} +{"episode_index": 2591, "length": 50} +{"episode_index": 2592, "length": 34} +{"episode_index": 2593, "length": 45} +{"episode_index": 2594, "length": 46} +{"episode_index": 2595, "length": 47} +{"episode_index": 2596, "length": 45} +{"episode_index": 2597, "length": 44} +{"episode_index": 2598, "length": 31} +{"episode_index": 2599, "length": 45} +{"episode_index": 2600, "length": 44} +{"episode_index": 2601, "length": 45} +{"episode_index": 2602, "length": 45} +{"episode_index": 2603, "length": 19} +{"episode_index": 2604, "length": 44} +{"episode_index": 2605, "length": 22} +{"episode_index": 2606, "length": 45} +{"episode_index": 2607, "length": 45} +{"episode_index": 2608, "length": 45} +{"episode_index": 2609, "length": 44} +{"episode_index": 2610, "length": 44} +{"episode_index": 2611, "length": 44} +{"episode_index": 2612, "length": 45} +{"episode_index": 2613, "length": 45} +{"episode_index": 2614, "length": 19} +{"episode_index": 2615, "length": 20} +{"episode_index": 2616, "length": 45} +{"episode_index": 2617, "length": 47} +{"episode_index": 2618, "length": 45} +{"episode_index": 2619, "length": 45} +{"episode_index": 2620, "length": 43} +{"episode_index": 2621, "length": 46} +{"episode_index": 2622, "length": 45} +{"episode_index": 2623, "length": 19} +{"episode_index": 2624, "length": 47} +{"episode_index": 2625, "length": 44} +{"episode_index": 2626, "length": 46} +{"episode_index": 2627, "length": 45} +{"episode_index": 2628, "length": 44} +{"episode_index": 2629, "length": 44} +{"episode_index": 2630, "length": 46} +{"episode_index": 2631, "length": 45} +{"episode_index": 2632, "length": 20} +{"episode_index": 2633, "length": 18} +{"episode_index": 2634, "length": 33} +{"episode_index": 2635, "length": 44} +{"episode_index": 2636, "length": 46} +{"episode_index": 2637, "length": 45} +{"episode_index": 2638, "length": 11} +{"episode_index": 2639, "length": 45} +{"episode_index": 2640, "length": 49} +{"episode_index": 2641, "length": 45} +{"episode_index": 2642, "length": 45} +{"episode_index": 2643, "length": 45} +{"episode_index": 2644, "length": 45} +{"episode_index": 2645, "length": 44} +{"episode_index": 2646, "length": 45} +{"episode_index": 2647, "length": 45} +{"episode_index": 2648, "length": 45} +{"episode_index": 2649, "length": 45} +{"episode_index": 2650, "length": 45} +{"episode_index": 2651, "length": 11} +{"episode_index": 2652, "length": 48} +{"episode_index": 2653, "length": 45} +{"episode_index": 2654, "length": 44} +{"episode_index": 2655, "length": 43} +{"episode_index": 2656, "length": 45} +{"episode_index": 2657, "length": 11} +{"episode_index": 2658, "length": 46} +{"episode_index": 2659, "length": 43} +{"episode_index": 2660, "length": 44} +{"episode_index": 2661, "length": 49} +{"episode_index": 2662, "length": 46} +{"episode_index": 2663, "length": 20} +{"episode_index": 2664, "length": 46} +{"episode_index": 2665, "length": 46} +{"episode_index": 2666, "length": 45} +{"episode_index": 2667, "length": 45} +{"episode_index": 2668, "length": 12} +{"episode_index": 2669, "length": 46} +{"episode_index": 2670, "length": 45} +{"episode_index": 2671, "length": 44} +{"episode_index": 2672, "length": 44} +{"episode_index": 2673, "length": 45} +{"episode_index": 2674, "length": 45} +{"episode_index": 2675, "length": 44} +{"episode_index": 2676, "length": 44} +{"episode_index": 2677, "length": 45} +{"episode_index": 2678, "length": 46} +{"episode_index": 2679, "length": 44} +{"episode_index": 2680, "length": 45} +{"episode_index": 2681, "length": 45} +{"episode_index": 2682, "length": 43} +{"episode_index": 2683, "length": 47} +{"episode_index": 2684, "length": 19} +{"episode_index": 2685, "length": 46} +{"episode_index": 2686, "length": 46} +{"episode_index": 2687, "length": 46} +{"episode_index": 2688, "length": 45} +{"episode_index": 2689, "length": 46} +{"episode_index": 2690, "length": 45} +{"episode_index": 2691, "length": 33} +{"episode_index": 2692, "length": 46} +{"episode_index": 2693, "length": 43} +{"episode_index": 2694, "length": 45} +{"episode_index": 2695, "length": 19} +{"episode_index": 2696, "length": 45} +{"episode_index": 2697, "length": 44} +{"episode_index": 2698, "length": 32} +{"episode_index": 2699, "length": 44} +{"episode_index": 2700, "length": 45} +{"episode_index": 2701, "length": 44} +{"episode_index": 2702, "length": 48} +{"episode_index": 2703, "length": 46} +{"episode_index": 2704, "length": 43} +{"episode_index": 2705, "length": 49} +{"episode_index": 2706, "length": 44} +{"episode_index": 2707, "length": 43} +{"episode_index": 2708, "length": 44} +{"episode_index": 2709, "length": 46} +{"episode_index": 2710, "length": 46} +{"episode_index": 2711, "length": 32} +{"episode_index": 2712, "length": 45} +{"episode_index": 2713, "length": 45} +{"episode_index": 2714, "length": 46} +{"episode_index": 2715, "length": 45} +{"episode_index": 2716, "length": 19} +{"episode_index": 2717, "length": 11} +{"episode_index": 2718, "length": 44} +{"episode_index": 2719, "length": 43} +{"episode_index": 2720, "length": 44} +{"episode_index": 2721, "length": 49} +{"episode_index": 2722, "length": 43} +{"episode_index": 2723, "length": 46} +{"episode_index": 2724, "length": 11} +{"episode_index": 2725, "length": 45} +{"episode_index": 2726, "length": 46} +{"episode_index": 2727, "length": 43} +{"episode_index": 2728, "length": 43} +{"episode_index": 2729, "length": 11} +{"episode_index": 2730, "length": 18} +{"episode_index": 2731, "length": 13} +{"episode_index": 2732, "length": 45} +{"episode_index": 2733, "length": 44} +{"episode_index": 2734, "length": 19} +{"episode_index": 2735, "length": 45} +{"episode_index": 2736, "length": 26} +{"episode_index": 2737, "length": 45} +{"episode_index": 2738, "length": 44} +{"episode_index": 2739, "length": 45} +{"episode_index": 2740, "length": 44} +{"episode_index": 2741, "length": 45} +{"episode_index": 2742, "length": 44} +{"episode_index": 2743, "length": 45} +{"episode_index": 2744, "length": 46} +{"episode_index": 2745, "length": 45} +{"episode_index": 2746, "length": 46} +{"episode_index": 2747, "length": 43} +{"episode_index": 2748, "length": 45} +{"episode_index": 2749, "length": 45} +{"episode_index": 2750, "length": 44} +{"episode_index": 2751, "length": 45} +{"episode_index": 2752, "length": 45} +{"episode_index": 2753, "length": 45} +{"episode_index": 2754, "length": 45} +{"episode_index": 2755, "length": 45} +{"episode_index": 2756, "length": 45} +{"episode_index": 2757, "length": 46} +{"episode_index": 2758, "length": 18} +{"episode_index": 2759, "length": 19} +{"episode_index": 2760, "length": 45} +{"episode_index": 2761, "length": 45} +{"episode_index": 2762, "length": 45} +{"episode_index": 2763, "length": 45} +{"episode_index": 2764, "length": 47} +{"episode_index": 2765, "length": 45} +{"episode_index": 2766, "length": 47} +{"episode_index": 2767, "length": 44} +{"episode_index": 2768, "length": 45} +{"episode_index": 2769, "length": 45} +{"episode_index": 2770, "length": 23} +{"episode_index": 2771, "length": 44} +{"episode_index": 2772, "length": 11} +{"episode_index": 2773, "length": 45} +{"episode_index": 2774, "length": 47} +{"episode_index": 2775, "length": 45} +{"episode_index": 2776, "length": 43} +{"episode_index": 2777, "length": 34} +{"episode_index": 2778, "length": 45} +{"episode_index": 2779, "length": 47} +{"episode_index": 2780, "length": 47} +{"episode_index": 2781, "length": 45} +{"episode_index": 2782, "length": 45} +{"episode_index": 2783, "length": 33} +{"episode_index": 2784, "length": 45} +{"episode_index": 2785, "length": 45} +{"episode_index": 2786, "length": 46} +{"episode_index": 2787, "length": 49} +{"episode_index": 2788, "length": 50} +{"episode_index": 2789, "length": 19} +{"episode_index": 2790, "length": 45} +{"episode_index": 2791, "length": 44} +{"episode_index": 2792, "length": 46} +{"episode_index": 2793, "length": 22} +{"episode_index": 2794, "length": 45} +{"episode_index": 2795, "length": 45} +{"episode_index": 2796, "length": 45} +{"episode_index": 2797, "length": 44} +{"episode_index": 2798, "length": 45} +{"episode_index": 2799, "length": 45} +{"episode_index": 2800, "length": 19} +{"episode_index": 2801, "length": 45} +{"episode_index": 2802, "length": 45} +{"episode_index": 2803, "length": 19} +{"episode_index": 2804, "length": 44} +{"episode_index": 2805, "length": 45} +{"episode_index": 2806, "length": 47} +{"episode_index": 2807, "length": 25} +{"episode_index": 2808, "length": 46} +{"episode_index": 2809, "length": 46} +{"episode_index": 2810, "length": 44} +{"episode_index": 2811, "length": 45} +{"episode_index": 2812, "length": 19} +{"episode_index": 2813, "length": 46} +{"episode_index": 2814, "length": 46} +{"episode_index": 2815, "length": 45} +{"episode_index": 2816, "length": 24} +{"episode_index": 2817, "length": 45} +{"episode_index": 2818, "length": 44} +{"episode_index": 2819, "length": 44} +{"episode_index": 2820, "length": 45} +{"episode_index": 2821, "length": 47} +{"episode_index": 2822, "length": 45} +{"episode_index": 2823, "length": 44} +{"episode_index": 2824, "length": 46} +{"episode_index": 2825, "length": 45} +{"episode_index": 2826, "length": 46} +{"episode_index": 2827, "length": 45} +{"episode_index": 2828, "length": 45} +{"episode_index": 2829, "length": 4} +{"episode_index": 2830, "length": 46} +{"episode_index": 2831, "length": 43} +{"episode_index": 2832, "length": 45} +{"episode_index": 2833, "length": 44} +{"episode_index": 2834, "length": 11} +{"episode_index": 2835, "length": 44} +{"episode_index": 2836, "length": 46} +{"episode_index": 2837, "length": 46} +{"episode_index": 2838, "length": 45} +{"episode_index": 2839, "length": 46} +{"episode_index": 2840, "length": 43} +{"episode_index": 2841, "length": 48} +{"episode_index": 2842, "length": 19} +{"episode_index": 2843, "length": 46} +{"episode_index": 2844, "length": 20} +{"episode_index": 2845, "length": 49} +{"episode_index": 2846, "length": 45} +{"episode_index": 2847, "length": 45} +{"episode_index": 2848, "length": 46} +{"episode_index": 2849, "length": 44} +{"episode_index": 2850, "length": 45} +{"episode_index": 2851, "length": 46} +{"episode_index": 2852, "length": 46} +{"episode_index": 2853, "length": 45} +{"episode_index": 2854, "length": 45} +{"episode_index": 2855, "length": 12} +{"episode_index": 2856, "length": 45} +{"episode_index": 2857, "length": 45} +{"episode_index": 2858, "length": 45} +{"episode_index": 2859, "length": 45} +{"episode_index": 2860, "length": 44} +{"episode_index": 2861, "length": 44} +{"episode_index": 2862, "length": 45} +{"episode_index": 2863, "length": 43} +{"episode_index": 2864, "length": 46} +{"episode_index": 2865, "length": 44} +{"episode_index": 2866, "length": 19} +{"episode_index": 2867, "length": 45} +{"episode_index": 2868, "length": 45} +{"episode_index": 2869, "length": 45} +{"episode_index": 2870, "length": 45} +{"episode_index": 2871, "length": 45} +{"episode_index": 2872, "length": 45} +{"episode_index": 2873, "length": 47} +{"episode_index": 2874, "length": 44} +{"episode_index": 2875, "length": 45} +{"episode_index": 2876, "length": 45} +{"episode_index": 2877, "length": 46} +{"episode_index": 2878, "length": 44} +{"episode_index": 2879, "length": 45} +{"episode_index": 2880, "length": 45} +{"episode_index": 2881, "length": 45} +{"episode_index": 2882, "length": 15} +{"episode_index": 2883, "length": 45} +{"episode_index": 2884, "length": 45} +{"episode_index": 2885, "length": 45} +{"episode_index": 2886, "length": 43} +{"episode_index": 2887, "length": 46} +{"episode_index": 2888, "length": 45} +{"episode_index": 2889, "length": 45} +{"episode_index": 2890, "length": 45} +{"episode_index": 2891, "length": 43} +{"episode_index": 2892, "length": 45} +{"episode_index": 2893, "length": 46} +{"episode_index": 2894, "length": 43} +{"episode_index": 2895, "length": 43} +{"episode_index": 2896, "length": 45} +{"episode_index": 2897, "length": 43} +{"episode_index": 2898, "length": 45} +{"episode_index": 2899, "length": 44} +{"episode_index": 2900, "length": 45} +{"episode_index": 2901, "length": 45} +{"episode_index": 2902, "length": 45} +{"episode_index": 2903, "length": 46} +{"episode_index": 2904, "length": 44} +{"episode_index": 2905, "length": 46} +{"episode_index": 2906, "length": 44} +{"episode_index": 2907, "length": 43} +{"episode_index": 2908, "length": 44} +{"episode_index": 2909, "length": 44} +{"episode_index": 2910, "length": 47} +{"episode_index": 2911, "length": 45} +{"episode_index": 2912, "length": 46} +{"episode_index": 2913, "length": 47} +{"episode_index": 2914, "length": 45} +{"episode_index": 2915, "length": 19} +{"episode_index": 2916, "length": 45} +{"episode_index": 2917, "length": 42} +{"episode_index": 2918, "length": 19} +{"episode_index": 2919, "length": 45} +{"episode_index": 2920, "length": 20} +{"episode_index": 2921, "length": 45} +{"episode_index": 2922, "length": 45} +{"episode_index": 2923, "length": 44} +{"episode_index": 2924, "length": 44} +{"episode_index": 2925, "length": 45} +{"episode_index": 2926, "length": 45} +{"episode_index": 2927, "length": 45} +{"episode_index": 2928, "length": 44} +{"episode_index": 2929, "length": 32} +{"episode_index": 2930, "length": 45} +{"episode_index": 2931, "length": 20} +{"episode_index": 2932, "length": 22} +{"episode_index": 2933, "length": 47} +{"episode_index": 2934, "length": 45} +{"episode_index": 2935, "length": 19} +{"episode_index": 2936, "length": 49} +{"episode_index": 2937, "length": 44} +{"episode_index": 2938, "length": 43} +{"episode_index": 2939, "length": 45} +{"episode_index": 2940, "length": 45} +{"episode_index": 2941, "length": 46} +{"episode_index": 2942, "length": 45} +{"episode_index": 2943, "length": 45} +{"episode_index": 2944, "length": 48} +{"episode_index": 2945, "length": 44} +{"episode_index": 2946, "length": 45} +{"episode_index": 2947, "length": 44} +{"episode_index": 2948, "length": 45} +{"episode_index": 2949, "length": 45} +{"episode_index": 2950, "length": 30} +{"episode_index": 2951, "length": 44} +{"episode_index": 2952, "length": 43} +{"episode_index": 2953, "length": 11} +{"episode_index": 2954, "length": 20} +{"episode_index": 2955, "length": 49} +{"episode_index": 2956, "length": 46} +{"episode_index": 2957, "length": 46} +{"episode_index": 2958, "length": 46} +{"episode_index": 2959, "length": 44} +{"episode_index": 2960, "length": 46} +{"episode_index": 2961, "length": 16} +{"episode_index": 2962, "length": 19} +{"episode_index": 2963, "length": 45} +{"episode_index": 2964, "length": 45} +{"episode_index": 2965, "length": 45} +{"episode_index": 2966, "length": 44} +{"episode_index": 2967, "length": 43} +{"episode_index": 2968, "length": 47} +{"episode_index": 2969, "length": 45} +{"episode_index": 2970, "length": 44} +{"episode_index": 2971, "length": 11} +{"episode_index": 2972, "length": 43} +{"episode_index": 2973, "length": 45} +{"episode_index": 2974, "length": 20} +{"episode_index": 2975, "length": 45} +{"episode_index": 2976, "length": 45} +{"episode_index": 2977, "length": 48} +{"episode_index": 2978, "length": 45} +{"episode_index": 2979, "length": 47} +{"episode_index": 2980, "length": 45} +{"episode_index": 2981, "length": 45} +{"episode_index": 2982, "length": 46} +{"episode_index": 2983, "length": 44} +{"episode_index": 2984, "length": 45} +{"episode_index": 2985, "length": 10} +{"episode_index": 2986, "length": 46} +{"episode_index": 2987, "length": 45} +{"episode_index": 2988, "length": 20} +{"episode_index": 2989, "length": 45} +{"episode_index": 2990, "length": 45} +{"episode_index": 2991, "length": 19} +{"episode_index": 2992, "length": 45} +{"episode_index": 2993, "length": 46} +{"episode_index": 2994, "length": 45} +{"episode_index": 2995, "length": 47} +{"episode_index": 2996, "length": 44} +{"episode_index": 2997, "length": 44} +{"episode_index": 2998, "length": 46} +{"episode_index": 2999, "length": 45} +{"episode_index": 3000, "length": 45} +{"episode_index": 3001, "length": 19} +{"episode_index": 3002, "length": 49} +{"episode_index": 3003, "length": 45} +{"episode_index": 3004, "length": 47} +{"episode_index": 3005, "length": 40} +{"episode_index": 3006, "length": 44} +{"episode_index": 3007, "length": 46} +{"episode_index": 3008, "length": 44} +{"episode_index": 3009, "length": 12} +{"episode_index": 3010, "length": 46} +{"episode_index": 3011, "length": 19} +{"episode_index": 3012, "length": 45} +{"episode_index": 3013, "length": 24} +{"episode_index": 3014, "length": 44} +{"episode_index": 3015, "length": 43} +{"episode_index": 3016, "length": 18} +{"episode_index": 3017, "length": 20} +{"episode_index": 3018, "length": 44} +{"episode_index": 3019, "length": 46} +{"episode_index": 3020, "length": 17} +{"episode_index": 3021, "length": 18} +{"episode_index": 3022, "length": 46} +{"episode_index": 3023, "length": 43} +{"episode_index": 3024, "length": 46} +{"episode_index": 3025, "length": 46} +{"episode_index": 3026, "length": 45} +{"episode_index": 3027, "length": 45} +{"episode_index": 3028, "length": 44} +{"episode_index": 3029, "length": 46} +{"episode_index": 3030, "length": 44} +{"episode_index": 3031, "length": 48} +{"episode_index": 3032, "length": 45} +{"episode_index": 3033, "length": 45} +{"episode_index": 3034, "length": 12} +{"episode_index": 3035, "length": 46} +{"episode_index": 3036, "length": 45} +{"episode_index": 3037, "length": 45} +{"episode_index": 3038, "length": 25} +{"episode_index": 3039, "length": 20} +{"episode_index": 3040, "length": 45} +{"episode_index": 3041, "length": 45} +{"episode_index": 3042, "length": 48} +{"episode_index": 3043, "length": 45} +{"episode_index": 3044, "length": 45} +{"episode_index": 3045, "length": 44} +{"episode_index": 3046, "length": 13} +{"episode_index": 3047, "length": 44} +{"episode_index": 3048, "length": 45} +{"episode_index": 3049, "length": 45} +{"episode_index": 3050, "length": 45} +{"episode_index": 3051, "length": 45} +{"episode_index": 3052, "length": 45} +{"episode_index": 3053, "length": 45} +{"episode_index": 3054, "length": 44} +{"episode_index": 3055, "length": 45} +{"episode_index": 3056, "length": 46} +{"episode_index": 3057, "length": 25} +{"episode_index": 3058, "length": 11} +{"episode_index": 3059, "length": 46} +{"episode_index": 3060, "length": 45} +{"episode_index": 3061, "length": 44} +{"episode_index": 3062, "length": 45} +{"episode_index": 3063, "length": 44} +{"episode_index": 3064, "length": 45} +{"episode_index": 3065, "length": 12} +{"episode_index": 3066, "length": 46} +{"episode_index": 3067, "length": 46} +{"episode_index": 3068, "length": 45} +{"episode_index": 3069, "length": 45} +{"episode_index": 3070, "length": 45} +{"episode_index": 3071, "length": 49} +{"episode_index": 3072, "length": 45} +{"episode_index": 3073, "length": 45} +{"episode_index": 3074, "length": 45} +{"episode_index": 3075, "length": 45} +{"episode_index": 3076, "length": 45} +{"episode_index": 3077, "length": 45} +{"episode_index": 3078, "length": 19} +{"episode_index": 3079, "length": 44} +{"episode_index": 3080, "length": 48} +{"episode_index": 3081, "length": 45} +{"episode_index": 3082, "length": 45} +{"episode_index": 3083, "length": 45} +{"episode_index": 3084, "length": 45} +{"episode_index": 3085, "length": 44} +{"episode_index": 3086, "length": 46} +{"episode_index": 3087, "length": 11} +{"episode_index": 3088, "length": 47} +{"episode_index": 3089, "length": 46} +{"episode_index": 3090, "length": 46} +{"episode_index": 3091, "length": 45} +{"episode_index": 3092, "length": 10} +{"episode_index": 3093, "length": 12} +{"episode_index": 3094, "length": 47} +{"episode_index": 3095, "length": 45} +{"episode_index": 3096, "length": 46} +{"episode_index": 3097, "length": 45} +{"episode_index": 3098, "length": 45} +{"episode_index": 3099, "length": 45} +{"episode_index": 3100, "length": 44} +{"episode_index": 3101, "length": 43} +{"episode_index": 3102, "length": 49} +{"episode_index": 3103, "length": 30} +{"episode_index": 3104, "length": 46} +{"episode_index": 3105, "length": 45} +{"episode_index": 3106, "length": 45} +{"episode_index": 3107, "length": 45} +{"episode_index": 3108, "length": 45} +{"episode_index": 3109, "length": 46} +{"episode_index": 3110, "length": 49} +{"episode_index": 3111, "length": 45} +{"episode_index": 3112, "length": 45} +{"episode_index": 3113, "length": 47} +{"episode_index": 3114, "length": 45} +{"episode_index": 3115, "length": 45} +{"episode_index": 3116, "length": 18} +{"episode_index": 3117, "length": 43} +{"episode_index": 3118, "length": 44} +{"episode_index": 3119, "length": 43} +{"episode_index": 3120, "length": 18} +{"episode_index": 3121, "length": 44} +{"episode_index": 3122, "length": 45} +{"episode_index": 3123, "length": 12} +{"episode_index": 3124, "length": 48} +{"episode_index": 3125, "length": 44} +{"episode_index": 3126, "length": 43} +{"episode_index": 3127, "length": 44} +{"episode_index": 3128, "length": 45} +{"episode_index": 3129, "length": 45} +{"episode_index": 3130, "length": 47} +{"episode_index": 3131, "length": 45} +{"episode_index": 3132, "length": 44} +{"episode_index": 3133, "length": 44} +{"episode_index": 3134, "length": 47} +{"episode_index": 3135, "length": 45} +{"episode_index": 3136, "length": 29} +{"episode_index": 3137, "length": 43} +{"episode_index": 3138, "length": 45} +{"episode_index": 3139, "length": 47} +{"episode_index": 3140, "length": 43} +{"episode_index": 3141, "length": 44} +{"episode_index": 3142, "length": 45} +{"episode_index": 3143, "length": 45} +{"episode_index": 3144, "length": 27} +{"episode_index": 3145, "length": 30} +{"episode_index": 3146, "length": 44} +{"episode_index": 3147, "length": 45} +{"episode_index": 3148, "length": 47} +{"episode_index": 3149, "length": 46} +{"episode_index": 3150, "length": 12} +{"episode_index": 3151, "length": 45} +{"episode_index": 3152, "length": 45} +{"episode_index": 3153, "length": 45} +{"episode_index": 3154, "length": 45} +{"episode_index": 3155, "length": 45} +{"episode_index": 3156, "length": 44} +{"episode_index": 3157, "length": 44} +{"episode_index": 3158, "length": 45} +{"episode_index": 3159, "length": 19} +{"episode_index": 3160, "length": 44} +{"episode_index": 3161, "length": 46} +{"episode_index": 3162, "length": 45} +{"episode_index": 3163, "length": 44} +{"episode_index": 3164, "length": 10} +{"episode_index": 3165, "length": 47} +{"episode_index": 3166, "length": 45} +{"episode_index": 3167, "length": 44} +{"episode_index": 3168, "length": 44} +{"episode_index": 3169, "length": 19} +{"episode_index": 3170, "length": 45} +{"episode_index": 3171, "length": 45} +{"episode_index": 3172, "length": 46} +{"episode_index": 3173, "length": 47} +{"episode_index": 3174, "length": 45} +{"episode_index": 3175, "length": 47} +{"episode_index": 3176, "length": 45} +{"episode_index": 3177, "length": 45} +{"episode_index": 3178, "length": 44} +{"episode_index": 3179, "length": 44} +{"episode_index": 3180, "length": 22} +{"episode_index": 3181, "length": 44} +{"episode_index": 3182, "length": 44} +{"episode_index": 3183, "length": 45} +{"episode_index": 3184, "length": 43} +{"episode_index": 3185, "length": 45} +{"episode_index": 3186, "length": 45} +{"episode_index": 3187, "length": 45} +{"episode_index": 3188, "length": 44} +{"episode_index": 3189, "length": 44} +{"episode_index": 3190, "length": 45} +{"episode_index": 3191, "length": 47} +{"episode_index": 3192, "length": 45} +{"episode_index": 3193, "length": 45} +{"episode_index": 3194, "length": 33} +{"episode_index": 3195, "length": 43} +{"episode_index": 3196, "length": 45} +{"episode_index": 3197, "length": 45} +{"episode_index": 3198, "length": 44} +{"episode_index": 3199, "length": 45} +{"episode_index": 3200, "length": 18} +{"episode_index": 3201, "length": 43} +{"episode_index": 3202, "length": 4} +{"episode_index": 3203, "length": 44} +{"episode_index": 3204, "length": 30} +{"episode_index": 3205, "length": 45} +{"episode_index": 3206, "length": 46} +{"episode_index": 3207, "length": 44} +{"episode_index": 3208, "length": 44} +{"episode_index": 3209, "length": 45} +{"episode_index": 3210, "length": 45} +{"episode_index": 3211, "length": 11} +{"episode_index": 3212, "length": 44} +{"episode_index": 3213, "length": 15} +{"episode_index": 3214, "length": 44} +{"episode_index": 3215, "length": 45} +{"episode_index": 3216, "length": 45} +{"episode_index": 3217, "length": 45} +{"episode_index": 3218, "length": 45} +{"episode_index": 3219, "length": 45} +{"episode_index": 3220, "length": 45} +{"episode_index": 3221, "length": 47} +{"episode_index": 3222, "length": 43} +{"episode_index": 3223, "length": 45} +{"episode_index": 3224, "length": 45} +{"episode_index": 3225, "length": 43} +{"episode_index": 3226, "length": 45} +{"episode_index": 3227, "length": 14} +{"episode_index": 3228, "length": 45} +{"episode_index": 3229, "length": 10} +{"episode_index": 3230, "length": 45} +{"episode_index": 3231, "length": 47} +{"episode_index": 3232, "length": 45} +{"episode_index": 3233, "length": 19} +{"episode_index": 3234, "length": 45} +{"episode_index": 3235, "length": 19} +{"episode_index": 3236, "length": 46} +{"episode_index": 3237, "length": 45} +{"episode_index": 3238, "length": 45} +{"episode_index": 3239, "length": 44} +{"episode_index": 3240, "length": 45} +{"episode_index": 3241, "length": 45} +{"episode_index": 3242, "length": 47} +{"episode_index": 3243, "length": 45} +{"episode_index": 3244, "length": 45} +{"episode_index": 3245, "length": 49} +{"episode_index": 3246, "length": 46} +{"episode_index": 3247, "length": 45} +{"episode_index": 3248, "length": 45} +{"episode_index": 3249, "length": 45} +{"episode_index": 3250, "length": 45} +{"episode_index": 3251, "length": 45} +{"episode_index": 3252, "length": 45} +{"episode_index": 3253, "length": 13} +{"episode_index": 3254, "length": 44} +{"episode_index": 3255, "length": 45} +{"episode_index": 3256, "length": 45} +{"episode_index": 3257, "length": 46} +{"episode_index": 3258, "length": 45} +{"episode_index": 3259, "length": 46} +{"episode_index": 3260, "length": 45} +{"episode_index": 3261, "length": 45} +{"episode_index": 3262, "length": 45} +{"episode_index": 3263, "length": 46} +{"episode_index": 3264, "length": 44} +{"episode_index": 3265, "length": 45} +{"episode_index": 3266, "length": 45} +{"episode_index": 3267, "length": 48} +{"episode_index": 3268, "length": 44} +{"episode_index": 3269, "length": 45} +{"episode_index": 3270, "length": 45} +{"episode_index": 3271, "length": 46} +{"episode_index": 3272, "length": 45} +{"episode_index": 3273, "length": 45} +{"episode_index": 3274, "length": 47} +{"episode_index": 3275, "length": 44} +{"episode_index": 3276, "length": 45} +{"episode_index": 3277, "length": 45} +{"episode_index": 3278, "length": 43} +{"episode_index": 3279, "length": 43} +{"episode_index": 3280, "length": 46} +{"episode_index": 3281, "length": 47} +{"episode_index": 3282, "length": 45} +{"episode_index": 3283, "length": 19} +{"episode_index": 3284, "length": 45} +{"episode_index": 3285, "length": 44} +{"episode_index": 3286, "length": 45} +{"episode_index": 3287, "length": 46} +{"episode_index": 3288, "length": 44} +{"episode_index": 3289, "length": 46} +{"episode_index": 3290, "length": 45} +{"episode_index": 3291, "length": 45} +{"episode_index": 3292, "length": 45} +{"episode_index": 3293, "length": 49} +{"episode_index": 3294, "length": 45} +{"episode_index": 3295, "length": 44} +{"episode_index": 3296, "length": 26} +{"episode_index": 3297, "length": 43} +{"episode_index": 3298, "length": 44} +{"episode_index": 3299, "length": 44} +{"episode_index": 3300, "length": 20} +{"episode_index": 3301, "length": 45} +{"episode_index": 3302, "length": 46} +{"episode_index": 3303, "length": 45} +{"episode_index": 3304, "length": 45} +{"episode_index": 3305, "length": 45} +{"episode_index": 3306, "length": 44} +{"episode_index": 3307, "length": 43} +{"episode_index": 3308, "length": 45} +{"episode_index": 3309, "length": 45} +{"episode_index": 3310, "length": 45} +{"episode_index": 3311, "length": 44} +{"episode_index": 3312, "length": 12} +{"episode_index": 3313, "length": 44} +{"episode_index": 3314, "length": 45} +{"episode_index": 3315, "length": 19} +{"episode_index": 3316, "length": 44} +{"episode_index": 3317, "length": 19} +{"episode_index": 3318, "length": 45} +{"episode_index": 3319, "length": 46} +{"episode_index": 3320, "length": 45} +{"episode_index": 3321, "length": 19} +{"episode_index": 3322, "length": 45} +{"episode_index": 3323, "length": 12} +{"episode_index": 3324, "length": 4} +{"episode_index": 3325, "length": 45} +{"episode_index": 3326, "length": 45} +{"episode_index": 3327, "length": 4} +{"episode_index": 3328, "length": 45} +{"episode_index": 3329, "length": 45} +{"episode_index": 3330, "length": 44} +{"episode_index": 3331, "length": 44} +{"episode_index": 3332, "length": 45} +{"episode_index": 3333, "length": 44} +{"episode_index": 3334, "length": 45} +{"episode_index": 3335, "length": 46} +{"episode_index": 3336, "length": 11} +{"episode_index": 3337, "length": 45} +{"episode_index": 3338, "length": 44} +{"episode_index": 3339, "length": 45} +{"episode_index": 3340, "length": 44} +{"episode_index": 3341, "length": 45} +{"episode_index": 3342, "length": 48} +{"episode_index": 3343, "length": 46} +{"episode_index": 3344, "length": 24} +{"episode_index": 3345, "length": 44} +{"episode_index": 3346, "length": 45} +{"episode_index": 3347, "length": 46} +{"episode_index": 3348, "length": 49} +{"episode_index": 3349, "length": 46} +{"episode_index": 3350, "length": 47} +{"episode_index": 3351, "length": 46} +{"episode_index": 3352, "length": 45} +{"episode_index": 3353, "length": 21} +{"episode_index": 3354, "length": 45} +{"episode_index": 3355, "length": 45} +{"episode_index": 3356, "length": 46} +{"episode_index": 3357, "length": 30} +{"episode_index": 3358, "length": 44} +{"episode_index": 3359, "length": 45} +{"episode_index": 3360, "length": 44} +{"episode_index": 3361, "length": 48} +{"episode_index": 3362, "length": 45} +{"episode_index": 3363, "length": 47} +{"episode_index": 3364, "length": 44} +{"episode_index": 3365, "length": 44} +{"episode_index": 3366, "length": 48} +{"episode_index": 3367, "length": 43} +{"episode_index": 3368, "length": 46} +{"episode_index": 3369, "length": 43} +{"episode_index": 3370, "length": 44} +{"episode_index": 3371, "length": 23} +{"episode_index": 3372, "length": 45} +{"episode_index": 3373, "length": 23} +{"episode_index": 3374, "length": 46} +{"episode_index": 3375, "length": 47} +{"episode_index": 3376, "length": 45} +{"episode_index": 3377, "length": 45} +{"episode_index": 3378, "length": 47} +{"episode_index": 3379, "length": 47} +{"episode_index": 3380, "length": 44} +{"episode_index": 3381, "length": 45} +{"episode_index": 3382, "length": 45} +{"episode_index": 3383, "length": 45} +{"episode_index": 3384, "length": 46} +{"episode_index": 3385, "length": 45} +{"episode_index": 3386, "length": 45} +{"episode_index": 3387, "length": 47} +{"episode_index": 3388, "length": 46} +{"episode_index": 3389, "length": 45} +{"episode_index": 3390, "length": 45} +{"episode_index": 3391, "length": 43} +{"episode_index": 3392, "length": 10} +{"episode_index": 3393, "length": 45} +{"episode_index": 3394, "length": 10} +{"episode_index": 3395, "length": 45} +{"episode_index": 3396, "length": 46} +{"episode_index": 3397, "length": 43} +{"episode_index": 3398, "length": 3} +{"episode_index": 3399, "length": 45} +{"episode_index": 3400, "length": 45} +{"episode_index": 3401, "length": 19} +{"episode_index": 3402, "length": 45} +{"episode_index": 3403, "length": 23} +{"episode_index": 3404, "length": 44} +{"episode_index": 3405, "length": 45} +{"episode_index": 3406, "length": 45} +{"episode_index": 3407, "length": 47} +{"episode_index": 3408, "length": 24} +{"episode_index": 3409, "length": 20} +{"episode_index": 3410, "length": 24} +{"episode_index": 3411, "length": 45} +{"episode_index": 3412, "length": 45} +{"episode_index": 3413, "length": 44} +{"episode_index": 3414, "length": 44} +{"episode_index": 3415, "length": 43} +{"episode_index": 3416, "length": 45} +{"episode_index": 3417, "length": 45} +{"episode_index": 3418, "length": 44} +{"episode_index": 3419, "length": 12} +{"episode_index": 3420, "length": 45} +{"episode_index": 3421, "length": 33} +{"episode_index": 3422, "length": 46} +{"episode_index": 3423, "length": 44} +{"episode_index": 3424, "length": 45} +{"episode_index": 3425, "length": 45} +{"episode_index": 3426, "length": 45} +{"episode_index": 3427, "length": 49} +{"episode_index": 3428, "length": 45} +{"episode_index": 3429, "length": 45} +{"episode_index": 3430, "length": 46} +{"episode_index": 3431, "length": 45} +{"episode_index": 3432, "length": 44} +{"episode_index": 3433, "length": 44} +{"episode_index": 3434, "length": 45} +{"episode_index": 3435, "length": 45} +{"episode_index": 3436, "length": 46} +{"episode_index": 3437, "length": 44} +{"episode_index": 3438, "length": 45} +{"episode_index": 3439, "length": 47} +{"episode_index": 3440, "length": 46} +{"episode_index": 3441, "length": 45} +{"episode_index": 3442, "length": 44} +{"episode_index": 3443, "length": 23} +{"episode_index": 3444, "length": 45} +{"episode_index": 3445, "length": 44} +{"episode_index": 3446, "length": 44} +{"episode_index": 3447, "length": 45} +{"episode_index": 3448, "length": 45} +{"episode_index": 3449, "length": 44} +{"episode_index": 3450, "length": 45} +{"episode_index": 3451, "length": 43} +{"episode_index": 3452, "length": 45} +{"episode_index": 3453, "length": 19} +{"episode_index": 3454, "length": 46} +{"episode_index": 3455, "length": 46} +{"episode_index": 3456, "length": 44} +{"episode_index": 3457, "length": 43} +{"episode_index": 3458, "length": 45} +{"episode_index": 3459, "length": 32} +{"episode_index": 3460, "length": 45} +{"episode_index": 3461, "length": 45} +{"episode_index": 3462, "length": 44} +{"episode_index": 3463, "length": 44} +{"episode_index": 3464, "length": 16} +{"episode_index": 3465, "length": 45} +{"episode_index": 3466, "length": 45} +{"episode_index": 3467, "length": 51} +{"episode_index": 3468, "length": 45} +{"episode_index": 3469, "length": 45} +{"episode_index": 3470, "length": 45} +{"episode_index": 3471, "length": 45} +{"episode_index": 3472, "length": 45} +{"episode_index": 3473, "length": 49} +{"episode_index": 3474, "length": 18} +{"episode_index": 3475, "length": 44} +{"episode_index": 3476, "length": 43} +{"episode_index": 3477, "length": 33} +{"episode_index": 3478, "length": 43} +{"episode_index": 3479, "length": 44} +{"episode_index": 3480, "length": 45} +{"episode_index": 3481, "length": 43} +{"episode_index": 3482, "length": 44} +{"episode_index": 3483, "length": 43} +{"episode_index": 3484, "length": 45} +{"episode_index": 3485, "length": 47} +{"episode_index": 3486, "length": 44} +{"episode_index": 3487, "length": 45} +{"episode_index": 3488, "length": 46} +{"episode_index": 3489, "length": 49} +{"episode_index": 3490, "length": 44} +{"episode_index": 3491, "length": 46} +{"episode_index": 3492, "length": 44} +{"episode_index": 3493, "length": 18} +{"episode_index": 3494, "length": 19} +{"episode_index": 3495, "length": 45} +{"episode_index": 3496, "length": 44} +{"episode_index": 3497, "length": 45} +{"episode_index": 3498, "length": 46} +{"episode_index": 3499, "length": 45} +{"episode_index": 3500, "length": 45} +{"episode_index": 3501, "length": 49} +{"episode_index": 3502, "length": 46} +{"episode_index": 3503, "length": 44} +{"episode_index": 3504, "length": 44} +{"episode_index": 3505, "length": 45} +{"episode_index": 3506, "length": 45} +{"episode_index": 3507, "length": 45} +{"episode_index": 3508, "length": 19} +{"episode_index": 3509, "length": 44} +{"episode_index": 3510, "length": 45} +{"episode_index": 3511, "length": 46} +{"episode_index": 3512, "length": 44} +{"episode_index": 3513, "length": 44} +{"episode_index": 3514, "length": 45} +{"episode_index": 3515, "length": 44} +{"episode_index": 3516, "length": 18} +{"episode_index": 3517, "length": 43} +{"episode_index": 3518, "length": 47} +{"episode_index": 3519, "length": 45} +{"episode_index": 3520, "length": 43} +{"episode_index": 3521, "length": 43} +{"episode_index": 3522, "length": 45} +{"episode_index": 3523, "length": 46} +{"episode_index": 3524, "length": 45} +{"episode_index": 3525, "length": 19} +{"episode_index": 3526, "length": 44} +{"episode_index": 3527, "length": 45} +{"episode_index": 3528, "length": 46} +{"episode_index": 3529, "length": 49} +{"episode_index": 3530, "length": 43} +{"episode_index": 3531, "length": 23} +{"episode_index": 3532, "length": 46} +{"episode_index": 3533, "length": 45} +{"episode_index": 3534, "length": 36} +{"episode_index": 3535, "length": 44} +{"episode_index": 3536, "length": 47} +{"episode_index": 3537, "length": 45} +{"episode_index": 3538, "length": 45} +{"episode_index": 3539, "length": 17} +{"episode_index": 3540, "length": 44} +{"episode_index": 3541, "length": 43} +{"episode_index": 3542, "length": 45} +{"episode_index": 3543, "length": 45} +{"episode_index": 3544, "length": 19} +{"episode_index": 3545, "length": 44} +{"episode_index": 3546, "length": 48} +{"episode_index": 3547, "length": 45} +{"episode_index": 3548, "length": 47} +{"episode_index": 3549, "length": 45} +{"episode_index": 3550, "length": 47} +{"episode_index": 3551, "length": 12} +{"episode_index": 3552, "length": 43} +{"episode_index": 3553, "length": 45} +{"episode_index": 3554, "length": 46} +{"episode_index": 3555, "length": 12} +{"episode_index": 3556, "length": 43} +{"episode_index": 3557, "length": 44} +{"episode_index": 3558, "length": 45} +{"episode_index": 3559, "length": 49} +{"episode_index": 3560, "length": 46} +{"episode_index": 3561, "length": 45} +{"episode_index": 3562, "length": 45} +{"episode_index": 3563, "length": 45} +{"episode_index": 3564, "length": 45} +{"episode_index": 3565, "length": 34} +{"episode_index": 3566, "length": 46} +{"episode_index": 3567, "length": 45} +{"episode_index": 3568, "length": 16} +{"episode_index": 3569, "length": 45} +{"episode_index": 3570, "length": 44} +{"episode_index": 3571, "length": 25} +{"episode_index": 3572, "length": 46} +{"episode_index": 3573, "length": 45} +{"episode_index": 3574, "length": 46} +{"episode_index": 3575, "length": 12} +{"episode_index": 3576, "length": 46} +{"episode_index": 3577, "length": 45} +{"episode_index": 3578, "length": 45} +{"episode_index": 3579, "length": 46} +{"episode_index": 3580, "length": 30} +{"episode_index": 3581, "length": 45} +{"episode_index": 3582, "length": 17} +{"episode_index": 3583, "length": 45} +{"episode_index": 3584, "length": 45} +{"episode_index": 3585, "length": 45} +{"episode_index": 3586, "length": 45} +{"episode_index": 3587, "length": 45} +{"episode_index": 3588, "length": 45} +{"episode_index": 3589, "length": 45} +{"episode_index": 3590, "length": 44} +{"episode_index": 3591, "length": 46} +{"episode_index": 3592, "length": 44} +{"episode_index": 3593, "length": 17} +{"episode_index": 3594, "length": 43} +{"episode_index": 3595, "length": 44} +{"episode_index": 3596, "length": 44} +{"episode_index": 3597, "length": 45} +{"episode_index": 3598, "length": 45} +{"episode_index": 3599, "length": 46} +{"episode_index": 3600, "length": 46} +{"episode_index": 3601, "length": 47} +{"episode_index": 3602, "length": 47} +{"episode_index": 3603, "length": 44} +{"episode_index": 3604, "length": 45} +{"episode_index": 3605, "length": 22} +{"episode_index": 3606, "length": 25} +{"episode_index": 3607, "length": 46} +{"episode_index": 3608, "length": 45} +{"episode_index": 3609, "length": 47} +{"episode_index": 3610, "length": 43} +{"episode_index": 3611, "length": 45} +{"episode_index": 3612, "length": 40} +{"episode_index": 3613, "length": 43} +{"episode_index": 3614, "length": 43} +{"episode_index": 3615, "length": 44} +{"episode_index": 3616, "length": 45} +{"episode_index": 3617, "length": 45} +{"episode_index": 3618, "length": 5} +{"episode_index": 3619, "length": 44} +{"episode_index": 3620, "length": 44} +{"episode_index": 3621, "length": 43} +{"episode_index": 3622, "length": 44} +{"episode_index": 3623, "length": 45} +{"episode_index": 3624, "length": 46} +{"episode_index": 3625, "length": 18} +{"episode_index": 3626, "length": 45} +{"episode_index": 3627, "length": 45} +{"episode_index": 3628, "length": 44} +{"episode_index": 3629, "length": 45} +{"episode_index": 3630, "length": 11} +{"episode_index": 3631, "length": 44} +{"episode_index": 3632, "length": 43} +{"episode_index": 3633, "length": 47} +{"episode_index": 3634, "length": 47} +{"episode_index": 3635, "length": 45} +{"episode_index": 3636, "length": 45} +{"episode_index": 3637, "length": 43} +{"episode_index": 3638, "length": 45} +{"episode_index": 3639, "length": 45} +{"episode_index": 3640, "length": 31} +{"episode_index": 3641, "length": 20} +{"episode_index": 3642, "length": 43} +{"episode_index": 3643, "length": 48} +{"episode_index": 3644, "length": 48} +{"episode_index": 3645, "length": 44} +{"episode_index": 3646, "length": 45} +{"episode_index": 3647, "length": 44} +{"episode_index": 3648, "length": 45} +{"episode_index": 3649, "length": 21} +{"episode_index": 3650, "length": 43} +{"episode_index": 3651, "length": 45} +{"episode_index": 3652, "length": 45} +{"episode_index": 3653, "length": 45} +{"episode_index": 3654, "length": 45} +{"episode_index": 3655, "length": 47} +{"episode_index": 3656, "length": 44} +{"episode_index": 3657, "length": 45} +{"episode_index": 3658, "length": 45} +{"episode_index": 3659, "length": 17} +{"episode_index": 3660, "length": 45} +{"episode_index": 3661, "length": 10} +{"episode_index": 3662, "length": 45} +{"episode_index": 3663, "length": 45} +{"episode_index": 3664, "length": 45} +{"episode_index": 3665, "length": 11} +{"episode_index": 3666, "length": 45} +{"episode_index": 3667, "length": 45} +{"episode_index": 3668, "length": 45} +{"episode_index": 3669, "length": 45} +{"episode_index": 3670, "length": 11} +{"episode_index": 3671, "length": 26} +{"episode_index": 3672, "length": 44} +{"episode_index": 3673, "length": 48} +{"episode_index": 3674, "length": 43} +{"episode_index": 3675, "length": 45} +{"episode_index": 3676, "length": 45} +{"episode_index": 3677, "length": 44} +{"episode_index": 3678, "length": 46} +{"episode_index": 3679, "length": 46} +{"episode_index": 3680, "length": 11} +{"episode_index": 3681, "length": 46} +{"episode_index": 3682, "length": 45} +{"episode_index": 3683, "length": 12} +{"episode_index": 3684, "length": 45} +{"episode_index": 3685, "length": 49} +{"episode_index": 3686, "length": 43} +{"episode_index": 3687, "length": 44} +{"episode_index": 3688, "length": 45} +{"episode_index": 3689, "length": 49} +{"episode_index": 3690, "length": 48} +{"episode_index": 3691, "length": 45} +{"episode_index": 3692, "length": 44} +{"episode_index": 3693, "length": 45} +{"episode_index": 3694, "length": 45} +{"episode_index": 3695, "length": 45} +{"episode_index": 3696, "length": 45} +{"episode_index": 3697, "length": 45} +{"episode_index": 3698, "length": 45} +{"episode_index": 3699, "length": 44} +{"episode_index": 3700, "length": 44} +{"episode_index": 3701, "length": 45} +{"episode_index": 3702, "length": 45} +{"episode_index": 3703, "length": 25} +{"episode_index": 3704, "length": 47} +{"episode_index": 3705, "length": 45} +{"episode_index": 3706, "length": 45} +{"episode_index": 3707, "length": 45} +{"episode_index": 3708, "length": 43} +{"episode_index": 3709, "length": 43} +{"episode_index": 3710, "length": 45} +{"episode_index": 3711, "length": 44} +{"episode_index": 3712, "length": 45} +{"episode_index": 3713, "length": 46} +{"episode_index": 3714, "length": 44} +{"episode_index": 3715, "length": 43} +{"episode_index": 3716, "length": 45} +{"episode_index": 3717, "length": 12} +{"episode_index": 3718, "length": 45} +{"episode_index": 3719, "length": 45} +{"episode_index": 3720, "length": 46} +{"episode_index": 3721, "length": 47} +{"episode_index": 3722, "length": 45} +{"episode_index": 3723, "length": 45} +{"episode_index": 3724, "length": 48} +{"episode_index": 3725, "length": 46} +{"episode_index": 3726, "length": 45} +{"episode_index": 3727, "length": 45} +{"episode_index": 3728, "length": 24} +{"episode_index": 3729, "length": 47} +{"episode_index": 3730, "length": 48} +{"episode_index": 3731, "length": 31} +{"episode_index": 3732, "length": 46} +{"episode_index": 3733, "length": 45} +{"episode_index": 3734, "length": 47} +{"episode_index": 3735, "length": 44} +{"episode_index": 3736, "length": 44} +{"episode_index": 3737, "length": 46} +{"episode_index": 3738, "length": 46} +{"episode_index": 3739, "length": 47} +{"episode_index": 3740, "length": 33} +{"episode_index": 3741, "length": 45} +{"episode_index": 3742, "length": 44} +{"episode_index": 3743, "length": 49} +{"episode_index": 3744, "length": 44} +{"episode_index": 3745, "length": 45} +{"episode_index": 3746, "length": 43} +{"episode_index": 3747, "length": 45} +{"episode_index": 3748, "length": 44} +{"episode_index": 3749, "length": 18} +{"episode_index": 3750, "length": 45} +{"episode_index": 3751, "length": 20} +{"episode_index": 3752, "length": 45} +{"episode_index": 3753, "length": 47} +{"episode_index": 3754, "length": 46} +{"episode_index": 3755, "length": 45} +{"episode_index": 3756, "length": 45} +{"episode_index": 3757, "length": 18} +{"episode_index": 3758, "length": 31} +{"episode_index": 3759, "length": 26} +{"episode_index": 3760, "length": 46} +{"episode_index": 3761, "length": 45} +{"episode_index": 3762, "length": 45} +{"episode_index": 3763, "length": 47} +{"episode_index": 3764, "length": 45} +{"episode_index": 3765, "length": 46} +{"episode_index": 3766, "length": 44} +{"episode_index": 3767, "length": 45} +{"episode_index": 3768, "length": 45} +{"episode_index": 3769, "length": 45} +{"episode_index": 3770, "length": 39} +{"episode_index": 3771, "length": 46} +{"episode_index": 3772, "length": 45} +{"episode_index": 3773, "length": 45} +{"episode_index": 3774, "length": 45} +{"episode_index": 3775, "length": 46} +{"episode_index": 3776, "length": 25} +{"episode_index": 3777, "length": 18} +{"episode_index": 3778, "length": 46} +{"episode_index": 3779, "length": 44} +{"episode_index": 3780, "length": 47} +{"episode_index": 3781, "length": 45} +{"episode_index": 3782, "length": 46} +{"episode_index": 3783, "length": 43} +{"episode_index": 3784, "length": 18} +{"episode_index": 3785, "length": 44} +{"episode_index": 3786, "length": 46} +{"episode_index": 3787, "length": 45} +{"episode_index": 3788, "length": 46} +{"episode_index": 3789, "length": 47} +{"episode_index": 3790, "length": 46} +{"episode_index": 3791, "length": 45} +{"episode_index": 3792, "length": 46} +{"episode_index": 3793, "length": 5} +{"episode_index": 3794, "length": 45} +{"episode_index": 3795, "length": 45} +{"episode_index": 3796, "length": 47} +{"episode_index": 3797, "length": 45} +{"episode_index": 3798, "length": 45} +{"episode_index": 3799, "length": 46} +{"episode_index": 3800, "length": 44} +{"episode_index": 3801, "length": 47} +{"episode_index": 3802, "length": 43} +{"episode_index": 3803, "length": 44} +{"episode_index": 3804, "length": 12} +{"episode_index": 3805, "length": 45} +{"episode_index": 3806, "length": 44} +{"episode_index": 3807, "length": 12} +{"episode_index": 3808, "length": 46} +{"episode_index": 3809, "length": 45} +{"episode_index": 3810, "length": 46} +{"episode_index": 3811, "length": 44} +{"episode_index": 3812, "length": 49} +{"episode_index": 3813, "length": 44} +{"episode_index": 3814, "length": 11} +{"episode_index": 3815, "length": 45} +{"episode_index": 3816, "length": 45} +{"episode_index": 3817, "length": 44} +{"episode_index": 3818, "length": 46} +{"episode_index": 3819, "length": 45} +{"episode_index": 3820, "length": 21} +{"episode_index": 3821, "length": 45} +{"episode_index": 3822, "length": 44} +{"episode_index": 3823, "length": 45} +{"episode_index": 3824, "length": 45} +{"episode_index": 3825, "length": 45} +{"episode_index": 3826, "length": 44} +{"episode_index": 3827, "length": 46} +{"episode_index": 3828, "length": 45} +{"episode_index": 3829, "length": 45} +{"episode_index": 3830, "length": 45} +{"episode_index": 3831, "length": 46} +{"episode_index": 3832, "length": 45} +{"episode_index": 3833, "length": 45} +{"episode_index": 3834, "length": 45} +{"episode_index": 3835, "length": 43} +{"episode_index": 3836, "length": 46} +{"episode_index": 3837, "length": 47} +{"episode_index": 3838, "length": 46} +{"episode_index": 3839, "length": 43} +{"episode_index": 3840, "length": 45} +{"episode_index": 3841, "length": 46} +{"episode_index": 3842, "length": 44} +{"episode_index": 3843, "length": 45} +{"episode_index": 3844, "length": 11} +{"episode_index": 3845, "length": 48} +{"episode_index": 3846, "length": 49} +{"episode_index": 3847, "length": 45} +{"episode_index": 3848, "length": 46} +{"episode_index": 3849, "length": 45} +{"episode_index": 3850, "length": 48} +{"episode_index": 3851, "length": 46} +{"episode_index": 3852, "length": 45} +{"episode_index": 3853, "length": 45} +{"episode_index": 3854, "length": 47} +{"episode_index": 3855, "length": 46} +{"episode_index": 3856, "length": 19} +{"episode_index": 3857, "length": 46} +{"episode_index": 3858, "length": 44} +{"episode_index": 3859, "length": 45} +{"episode_index": 3860, "length": 45} +{"episode_index": 3861, "length": 45} +{"episode_index": 3862, "length": 44} +{"episode_index": 3863, "length": 45} +{"episode_index": 3864, "length": 46} +{"episode_index": 3865, "length": 43} +{"episode_index": 3866, "length": 46} +{"episode_index": 3867, "length": 45} +{"episode_index": 3868, "length": 49} +{"episode_index": 3869, "length": 15} +{"episode_index": 3870, "length": 44} +{"episode_index": 3871, "length": 45} +{"episode_index": 3872, "length": 11} +{"episode_index": 3873, "length": 45} +{"episode_index": 3874, "length": 45} +{"episode_index": 3875, "length": 44} +{"episode_index": 3876, "length": 19} +{"episode_index": 3877, "length": 47} +{"episode_index": 3878, "length": 47} +{"episode_index": 3879, "length": 45} +{"episode_index": 3880, "length": 45} +{"episode_index": 3881, "length": 45} +{"episode_index": 3882, "length": 44} +{"episode_index": 3883, "length": 45} +{"episode_index": 3884, "length": 45} +{"episode_index": 3885, "length": 45} +{"episode_index": 3886, "length": 11} +{"episode_index": 3887, "length": 18} +{"episode_index": 3888, "length": 45} +{"episode_index": 3889, "length": 47} +{"episode_index": 3890, "length": 45} +{"episode_index": 3891, "length": 45} +{"episode_index": 3892, "length": 44} +{"episode_index": 3893, "length": 45} +{"episode_index": 3894, "length": 45} +{"episode_index": 3895, "length": 11} +{"episode_index": 3896, "length": 43} +{"episode_index": 3897, "length": 45} +{"episode_index": 3898, "length": 13} +{"episode_index": 3899, "length": 45} +{"episode_index": 3900, "length": 45} +{"episode_index": 3901, "length": 49} +{"episode_index": 3902, "length": 45} +{"episode_index": 3903, "length": 45} +{"episode_index": 3904, "length": 45} +{"episode_index": 3905, "length": 46} +{"episode_index": 3906, "length": 44} +{"episode_index": 3907, "length": 45} +{"episode_index": 3908, "length": 15} +{"episode_index": 3909, "length": 13} +{"episode_index": 3910, "length": 45} +{"episode_index": 3911, "length": 44} +{"episode_index": 3912, "length": 44} +{"episode_index": 3913, "length": 45} +{"episode_index": 3914, "length": 45} +{"episode_index": 3915, "length": 45} +{"episode_index": 3916, "length": 43} +{"episode_index": 3917, "length": 46} +{"episode_index": 3918, "length": 46} +{"episode_index": 3919, "length": 45} +{"episode_index": 3920, "length": 24} +{"episode_index": 3921, "length": 44} +{"episode_index": 3922, "length": 45} +{"episode_index": 3923, "length": 45} +{"episode_index": 3924, "length": 46} +{"episode_index": 3925, "length": 45} +{"episode_index": 3926, "length": 45} +{"episode_index": 3927, "length": 45} +{"episode_index": 3928, "length": 44} +{"episode_index": 3929, "length": 46} +{"episode_index": 3930, "length": 45} +{"episode_index": 3931, "length": 46} +{"episode_index": 3932, "length": 46} +{"episode_index": 3933, "length": 25} +{"episode_index": 3934, "length": 46} +{"episode_index": 3935, "length": 45} +{"episode_index": 3936, "length": 45} +{"episode_index": 3937, "length": 47} +{"episode_index": 3938, "length": 47} +{"episode_index": 3939, "length": 45} +{"episode_index": 3940, "length": 45} +{"episode_index": 3941, "length": 45} +{"episode_index": 3942, "length": 47} +{"episode_index": 3943, "length": 44} +{"episode_index": 3944, "length": 45} +{"episode_index": 3945, "length": 46} +{"episode_index": 3946, "length": 44} +{"episode_index": 3947, "length": 45} +{"episode_index": 3948, "length": 44} +{"episode_index": 3949, "length": 20} +{"episode_index": 3950, "length": 45} +{"episode_index": 3951, "length": 45} +{"episode_index": 3952, "length": 44} +{"episode_index": 3953, "length": 46} +{"episode_index": 3954, "length": 45} +{"episode_index": 3955, "length": 45} +{"episode_index": 3956, "length": 46} +{"episode_index": 3957, "length": 44} +{"episode_index": 3958, "length": 43} +{"episode_index": 3959, "length": 46} +{"episode_index": 3960, "length": 45} +{"episode_index": 3961, "length": 46} +{"episode_index": 3962, "length": 24} +{"episode_index": 3963, "length": 44} +{"episode_index": 3964, "length": 46} +{"episode_index": 3965, "length": 43} +{"episode_index": 3966, "length": 44} +{"episode_index": 3967, "length": 45} +{"episode_index": 3968, "length": 32} +{"episode_index": 3969, "length": 46} +{"episode_index": 3970, "length": 45} +{"episode_index": 3971, "length": 45} +{"episode_index": 3972, "length": 10} +{"episode_index": 3973, "length": 47} +{"episode_index": 3974, "length": 46} +{"episode_index": 3975, "length": 50} +{"episode_index": 3976, "length": 47} +{"episode_index": 3977, "length": 45} +{"episode_index": 3978, "length": 46} +{"episode_index": 3979, "length": 44} +{"episode_index": 3980, "length": 45} +{"episode_index": 3981, "length": 44} +{"episode_index": 3982, "length": 15} +{"episode_index": 3983, "length": 5} +{"episode_index": 3984, "length": 45} +{"episode_index": 3985, "length": 45} +{"episode_index": 3986, "length": 29} +{"episode_index": 3987, "length": 44} +{"episode_index": 3988, "length": 45} +{"episode_index": 3989, "length": 45} +{"episode_index": 3990, "length": 45} +{"episode_index": 3991, "length": 46} +{"episode_index": 3992, "length": 46} +{"episode_index": 3993, "length": 44} +{"episode_index": 3994, "length": 48} +{"episode_index": 3995, "length": 45} +{"episode_index": 3996, "length": 12} +{"episode_index": 3997, "length": 45} +{"episode_index": 3998, "length": 18} +{"episode_index": 3999, "length": 44} +{"episode_index": 4000, "length": 46} +{"episode_index": 4001, "length": 44} +{"episode_index": 4002, "length": 45} +{"episode_index": 4003, "length": 11} +{"episode_index": 4004, "length": 46} +{"episode_index": 4005, "length": 18} +{"episode_index": 4006, "length": 44} +{"episode_index": 4007, "length": 45} +{"episode_index": 4008, "length": 45} +{"episode_index": 4009, "length": 46} +{"episode_index": 4010, "length": 44} +{"episode_index": 4011, "length": 47} +{"episode_index": 4012, "length": 19} +{"episode_index": 4013, "length": 45} +{"episode_index": 4014, "length": 45} +{"episode_index": 4015, "length": 43} +{"episode_index": 4016, "length": 43} +{"episode_index": 4017, "length": 47} +{"episode_index": 4018, "length": 44} +{"episode_index": 4019, "length": 12} +{"episode_index": 4020, "length": 45} +{"episode_index": 4021, "length": 43} +{"episode_index": 4022, "length": 45} +{"episode_index": 4023, "length": 46} +{"episode_index": 4024, "length": 45} +{"episode_index": 4025, "length": 47} +{"episode_index": 4026, "length": 46} +{"episode_index": 4027, "length": 45} +{"episode_index": 4028, "length": 46} +{"episode_index": 4029, "length": 45} +{"episode_index": 4030, "length": 45} +{"episode_index": 4031, "length": 44} +{"episode_index": 4032, "length": 23} +{"episode_index": 4033, "length": 45} +{"episode_index": 4034, "length": 45} +{"episode_index": 4035, "length": 45} +{"episode_index": 4036, "length": 45} +{"episode_index": 4037, "length": 45} +{"episode_index": 4038, "length": 44} +{"episode_index": 4039, "length": 46} +{"episode_index": 4040, "length": 46} +{"episode_index": 4041, "length": 12} +{"episode_index": 4042, "length": 46} +{"episode_index": 4043, "length": 45} +{"episode_index": 4044, "length": 45} +{"episode_index": 4045, "length": 46} +{"episode_index": 4046, "length": 44} +{"episode_index": 4047, "length": 46} +{"episode_index": 4048, "length": 11} +{"episode_index": 4049, "length": 44} +{"episode_index": 4050, "length": 45} +{"episode_index": 4051, "length": 44} +{"episode_index": 4052, "length": 44} +{"episode_index": 4053, "length": 43} +{"episode_index": 4054, "length": 45} +{"episode_index": 4055, "length": 44} +{"episode_index": 4056, "length": 18} +{"episode_index": 4057, "length": 18} +{"episode_index": 4058, "length": 44} +{"episode_index": 4059, "length": 45} +{"episode_index": 4060, "length": 30} +{"episode_index": 4061, "length": 46} +{"episode_index": 4062, "length": 45} +{"episode_index": 4063, "length": 51} +{"episode_index": 4064, "length": 12} +{"episode_index": 4065, "length": 11} +{"episode_index": 4066, "length": 43} +{"episode_index": 4067, "length": 46} +{"episode_index": 4068, "length": 45} +{"episode_index": 4069, "length": 30} +{"episode_index": 4070, "length": 48} +{"episode_index": 4071, "length": 18} +{"episode_index": 4072, "length": 18} +{"episode_index": 4073, "length": 46} +{"episode_index": 4074, "length": 45} +{"episode_index": 4075, "length": 45} +{"episode_index": 4076, "length": 43} +{"episode_index": 4077, "length": 47} +{"episode_index": 4078, "length": 44} +{"episode_index": 4079, "length": 44} +{"episode_index": 4080, "length": 12} +{"episode_index": 4081, "length": 44} +{"episode_index": 4082, "length": 44} +{"episode_index": 4083, "length": 47} +{"episode_index": 4084, "length": 45} +{"episode_index": 4085, "length": 44} +{"episode_index": 4086, "length": 45} +{"episode_index": 4087, "length": 45} +{"episode_index": 4088, "length": 45} +{"episode_index": 4089, "length": 44} +{"episode_index": 4090, "length": 44} +{"episode_index": 4091, "length": 45} +{"episode_index": 4092, "length": 45} +{"episode_index": 4093, "length": 45} +{"episode_index": 4094, "length": 45} +{"episode_index": 4095, "length": 45} +{"episode_index": 4096, "length": 33} +{"episode_index": 4097, "length": 12} +{"episode_index": 4098, "length": 43} +{"episode_index": 4099, "length": 19} +{"episode_index": 4100, "length": 44} +{"episode_index": 4101, "length": 44} +{"episode_index": 4102, "length": 44} +{"episode_index": 4103, "length": 44} +{"episode_index": 4104, "length": 46} +{"episode_index": 4105, "length": 44} +{"episode_index": 4106, "length": 34} +{"episode_index": 4107, "length": 43} +{"episode_index": 4108, "length": 47} +{"episode_index": 4109, "length": 46} +{"episode_index": 4110, "length": 45} +{"episode_index": 4111, "length": 45} +{"episode_index": 4112, "length": 44} +{"episode_index": 4113, "length": 45} +{"episode_index": 4114, "length": 45} +{"episode_index": 4115, "length": 44} +{"episode_index": 4116, "length": 18} +{"episode_index": 4117, "length": 18} +{"episode_index": 4118, "length": 45} +{"episode_index": 4119, "length": 45} +{"episode_index": 4120, "length": 44} +{"episode_index": 4121, "length": 30} +{"episode_index": 4122, "length": 43} +{"episode_index": 4123, "length": 45} +{"episode_index": 4124, "length": 46} +{"episode_index": 4125, "length": 47} +{"episode_index": 4126, "length": 45} +{"episode_index": 4127, "length": 45} +{"episode_index": 4128, "length": 45} +{"episode_index": 4129, "length": 45} +{"episode_index": 4130, "length": 46} +{"episode_index": 4131, "length": 45} +{"episode_index": 4132, "length": 43} +{"episode_index": 4133, "length": 10} +{"episode_index": 4134, "length": 45} +{"episode_index": 4135, "length": 45} +{"episode_index": 4136, "length": 20} +{"episode_index": 4137, "length": 46} +{"episode_index": 4138, "length": 26} +{"episode_index": 4139, "length": 25} +{"episode_index": 4140, "length": 43} +{"episode_index": 4141, "length": 44} +{"episode_index": 4142, "length": 45} +{"episode_index": 4143, "length": 45} +{"episode_index": 4144, "length": 45} +{"episode_index": 4145, "length": 19} +{"episode_index": 4146, "length": 44} +{"episode_index": 4147, "length": 45} +{"episode_index": 4148, "length": 44} +{"episode_index": 4149, "length": 10} +{"episode_index": 4150, "length": 45} +{"episode_index": 4151, "length": 45} +{"episode_index": 4152, "length": 44} +{"episode_index": 4153, "length": 45} +{"episode_index": 4154, "length": 46} +{"episode_index": 4155, "length": 45} +{"episode_index": 4156, "length": 33} +{"episode_index": 4157, "length": 45} +{"episode_index": 4158, "length": 45} +{"episode_index": 4159, "length": 45} +{"episode_index": 4160, "length": 44} +{"episode_index": 4161, "length": 46} +{"episode_index": 4162, "length": 44} +{"episode_index": 4163, "length": 47} +{"episode_index": 4164, "length": 12} +{"episode_index": 4165, "length": 44} +{"episode_index": 4166, "length": 44} +{"episode_index": 4167, "length": 44} +{"episode_index": 4168, "length": 45} +{"episode_index": 4169, "length": 47} +{"episode_index": 4170, "length": 45} +{"episode_index": 4171, "length": 45} +{"episode_index": 4172, "length": 46} +{"episode_index": 4173, "length": 45} +{"episode_index": 4174, "length": 47} +{"episode_index": 4175, "length": 43} +{"episode_index": 4176, "length": 45} +{"episode_index": 4177, "length": 45} +{"episode_index": 4178, "length": 45} +{"episode_index": 4179, "length": 45} +{"episode_index": 4180, "length": 45} +{"episode_index": 4181, "length": 45} +{"episode_index": 4182, "length": 44} +{"episode_index": 4183, "length": 45} +{"episode_index": 4184, "length": 45} +{"episode_index": 4185, "length": 44} +{"episode_index": 4186, "length": 46} +{"episode_index": 4187, "length": 46} +{"episode_index": 4188, "length": 19} +{"episode_index": 4189, "length": 46} +{"episode_index": 4190, "length": 43} +{"episode_index": 4191, "length": 43} +{"episode_index": 4192, "length": 47} +{"episode_index": 4193, "length": 45} +{"episode_index": 4194, "length": 43} +{"episode_index": 4195, "length": 45} +{"episode_index": 4196, "length": 49} +{"episode_index": 4197, "length": 45} +{"episode_index": 4198, "length": 43} +{"episode_index": 4199, "length": 44} +{"episode_index": 4200, "length": 44} +{"episode_index": 4201, "length": 45} +{"episode_index": 4202, "length": 45} +{"episode_index": 4203, "length": 45} +{"episode_index": 4204, "length": 45} +{"episode_index": 4205, "length": 47} +{"episode_index": 4206, "length": 31} +{"episode_index": 4207, "length": 45} +{"episode_index": 4208, "length": 20} +{"episode_index": 4209, "length": 47} +{"episode_index": 4210, "length": 11} +{"episode_index": 4211, "length": 43} +{"episode_index": 4212, "length": 46} +{"episode_index": 4213, "length": 44} +{"episode_index": 4214, "length": 44} +{"episode_index": 4215, "length": 11} +{"episode_index": 4216, "length": 45} +{"episode_index": 4217, "length": 43} +{"episode_index": 4218, "length": 20} +{"episode_index": 4219, "length": 45} +{"episode_index": 4220, "length": 45} +{"episode_index": 4221, "length": 47} +{"episode_index": 4222, "length": 45} +{"episode_index": 4223, "length": 12} +{"episode_index": 4224, "length": 48} +{"episode_index": 4225, "length": 46} +{"episode_index": 4226, "length": 46} +{"episode_index": 4227, "length": 45} +{"episode_index": 4228, "length": 45} +{"episode_index": 4229, "length": 43} +{"episode_index": 4230, "length": 44} +{"episode_index": 4231, "length": 43} +{"episode_index": 4232, "length": 42} +{"episode_index": 4233, "length": 44} +{"episode_index": 4234, "length": 49} +{"episode_index": 4235, "length": 44} +{"episode_index": 4236, "length": 43} +{"episode_index": 4237, "length": 45} +{"episode_index": 4238, "length": 3} +{"episode_index": 4239, "length": 47} +{"episode_index": 4240, "length": 47} +{"episode_index": 4241, "length": 44} +{"episode_index": 4242, "length": 45} +{"episode_index": 4243, "length": 45} +{"episode_index": 4244, "length": 43} +{"episode_index": 4245, "length": 44} +{"episode_index": 4246, "length": 45} +{"episode_index": 4247, "length": 45} +{"episode_index": 4248, "length": 44} +{"episode_index": 4249, "length": 47} +{"episode_index": 4250, "length": 32} +{"episode_index": 4251, "length": 16} +{"episode_index": 4252, "length": 47} +{"episode_index": 4253, "length": 45} +{"episode_index": 4254, "length": 44} +{"episode_index": 4255, "length": 45} +{"episode_index": 4256, "length": 19} +{"episode_index": 4257, "length": 44} +{"episode_index": 4258, "length": 11} +{"episode_index": 4259, "length": 45} +{"episode_index": 4260, "length": 48} +{"episode_index": 4261, "length": 43} +{"episode_index": 4262, "length": 19} +{"episode_index": 4263, "length": 45} +{"episode_index": 4264, "length": 19} +{"episode_index": 4265, "length": 45} +{"episode_index": 4266, "length": 44} +{"episode_index": 4267, "length": 45} +{"episode_index": 4268, "length": 45} +{"episode_index": 4269, "length": 45} +{"episode_index": 4270, "length": 44} +{"episode_index": 4271, "length": 45} +{"episode_index": 4272, "length": 45} +{"episode_index": 4273, "length": 45} +{"episode_index": 4274, "length": 47} +{"episode_index": 4275, "length": 18} +{"episode_index": 4276, "length": 44} +{"episode_index": 4277, "length": 44} +{"episode_index": 4278, "length": 45} +{"episode_index": 4279, "length": 11} +{"episode_index": 4280, "length": 45} +{"episode_index": 4281, "length": 45} +{"episode_index": 4282, "length": 48} +{"episode_index": 4283, "length": 45} +{"episode_index": 4284, "length": 12} +{"episode_index": 4285, "length": 45} +{"episode_index": 4286, "length": 44} +{"episode_index": 4287, "length": 45} +{"episode_index": 4288, "length": 49} +{"episode_index": 4289, "length": 20} +{"episode_index": 4290, "length": 44} +{"episode_index": 4291, "length": 45} +{"episode_index": 4292, "length": 45} +{"episode_index": 4293, "length": 44} +{"episode_index": 4294, "length": 45} +{"episode_index": 4295, "length": 43} +{"episode_index": 4296, "length": 47} +{"episode_index": 4297, "length": 45} +{"episode_index": 4298, "length": 45} +{"episode_index": 4299, "length": 44} +{"episode_index": 4300, "length": 43} +{"episode_index": 4301, "length": 5} +{"episode_index": 4302, "length": 45} +{"episode_index": 4303, "length": 49} +{"episode_index": 4304, "length": 27} +{"episode_index": 4305, "length": 45} +{"episode_index": 4306, "length": 45} +{"episode_index": 4307, "length": 45} +{"episode_index": 4308, "length": 44} +{"episode_index": 4309, "length": 5} +{"episode_index": 4310, "length": 25} +{"episode_index": 4311, "length": 45} +{"episode_index": 4312, "length": 44} +{"episode_index": 4313, "length": 44} +{"episode_index": 4314, "length": 45} +{"episode_index": 4315, "length": 46} +{"episode_index": 4316, "length": 44} +{"episode_index": 4317, "length": 48} +{"episode_index": 4318, "length": 47} +{"episode_index": 4319, "length": 44} +{"episode_index": 4320, "length": 47} +{"episode_index": 4321, "length": 45} +{"episode_index": 4322, "length": 19} +{"episode_index": 4323, "length": 12} +{"episode_index": 4324, "length": 44} +{"episode_index": 4325, "length": 47} +{"episode_index": 4326, "length": 46} +{"episode_index": 4327, "length": 44} +{"episode_index": 4328, "length": 45} +{"episode_index": 4329, "length": 45} +{"episode_index": 4330, "length": 45} +{"episode_index": 4331, "length": 18} +{"episode_index": 4332, "length": 46} +{"episode_index": 4333, "length": 44} +{"episode_index": 4334, "length": 47} +{"episode_index": 4335, "length": 45} +{"episode_index": 4336, "length": 24} +{"episode_index": 4337, "length": 43} +{"episode_index": 4338, "length": 45} +{"episode_index": 4339, "length": 46} +{"episode_index": 4340, "length": 44} +{"episode_index": 4341, "length": 45} +{"episode_index": 4342, "length": 46} +{"episode_index": 4343, "length": 45} +{"episode_index": 4344, "length": 44} +{"episode_index": 4345, "length": 45} +{"episode_index": 4346, "length": 46} +{"episode_index": 4347, "length": 47} +{"episode_index": 4348, "length": 48} +{"episode_index": 4349, "length": 44} +{"episode_index": 4350, "length": 45} +{"episode_index": 4351, "length": 47} +{"episode_index": 4352, "length": 46} +{"episode_index": 4353, "length": 45} +{"episode_index": 4354, "length": 45} +{"episode_index": 4355, "length": 45} +{"episode_index": 4356, "length": 45} +{"episode_index": 4357, "length": 45} +{"episode_index": 4358, "length": 12} +{"episode_index": 4359, "length": 43} +{"episode_index": 4360, "length": 45} +{"episode_index": 4361, "length": 45} +{"episode_index": 4362, "length": 45} +{"episode_index": 4363, "length": 46} +{"episode_index": 4364, "length": 45} +{"episode_index": 4365, "length": 43} +{"episode_index": 4366, "length": 48} +{"episode_index": 4367, "length": 46} +{"episode_index": 4368, "length": 43} +{"episode_index": 4369, "length": 47} +{"episode_index": 4370, "length": 45} +{"episode_index": 4371, "length": 47} +{"episode_index": 4372, "length": 46} +{"episode_index": 4373, "length": 45} +{"episode_index": 4374, "length": 44} +{"episode_index": 4375, "length": 44} +{"episode_index": 4376, "length": 45} +{"episode_index": 4377, "length": 45} +{"episode_index": 4378, "length": 47} +{"episode_index": 4379, "length": 46} +{"episode_index": 4380, "length": 46} +{"episode_index": 4381, "length": 46} +{"episode_index": 4382, "length": 44} +{"episode_index": 4383, "length": 45} +{"episode_index": 4384, "length": 45} +{"episode_index": 4385, "length": 46} +{"episode_index": 4386, "length": 45} +{"episode_index": 4387, "length": 45} +{"episode_index": 4388, "length": 12} +{"episode_index": 4389, "length": 44} +{"episode_index": 4390, "length": 45} +{"episode_index": 4391, "length": 47} +{"episode_index": 4392, "length": 44} +{"episode_index": 4393, "length": 47} +{"episode_index": 4394, "length": 46} +{"episode_index": 4395, "length": 44} +{"episode_index": 4396, "length": 47} +{"episode_index": 4397, "length": 46} +{"episode_index": 4398, "length": 45} +{"episode_index": 4399, "length": 46} +{"episode_index": 4400, "length": 46} +{"episode_index": 4401, "length": 18} +{"episode_index": 4402, "length": 43} +{"episode_index": 4403, "length": 24} +{"episode_index": 4404, "length": 46} +{"episode_index": 4405, "length": 45} +{"episode_index": 4406, "length": 44} +{"episode_index": 4407, "length": 45} +{"episode_index": 4408, "length": 44} +{"episode_index": 4409, "length": 44} +{"episode_index": 4410, "length": 43} +{"episode_index": 4411, "length": 45} +{"episode_index": 4412, "length": 45} +{"episode_index": 4413, "length": 46} +{"episode_index": 4414, "length": 18} +{"episode_index": 4415, "length": 12} +{"episode_index": 4416, "length": 44} +{"episode_index": 4417, "length": 18} +{"episode_index": 4418, "length": 46} +{"episode_index": 4419, "length": 45} +{"episode_index": 4420, "length": 46} +{"episode_index": 4421, "length": 45} +{"episode_index": 4422, "length": 5} +{"episode_index": 4423, "length": 44} +{"episode_index": 4424, "length": 46} +{"episode_index": 4425, "length": 44} +{"episode_index": 4426, "length": 12} +{"episode_index": 4427, "length": 45} +{"episode_index": 4428, "length": 49} +{"episode_index": 4429, "length": 45} +{"episode_index": 4430, "length": 44} +{"episode_index": 4431, "length": 49} +{"episode_index": 4432, "length": 46} +{"episode_index": 4433, "length": 18} +{"episode_index": 4434, "length": 45} +{"episode_index": 4435, "length": 47} +{"episode_index": 4436, "length": 44} +{"episode_index": 4437, "length": 43} +{"episode_index": 4438, "length": 47} +{"episode_index": 4439, "length": 45} +{"episode_index": 4440, "length": 44} +{"episode_index": 4441, "length": 45} +{"episode_index": 4442, "length": 33} +{"episode_index": 4443, "length": 45} +{"episode_index": 4444, "length": 46} +{"episode_index": 4445, "length": 45} +{"episode_index": 4446, "length": 44} +{"episode_index": 4447, "length": 47} +{"episode_index": 4448, "length": 45} +{"episode_index": 4449, "length": 46} +{"episode_index": 4450, "length": 45} +{"episode_index": 4451, "length": 26} +{"episode_index": 4452, "length": 45} +{"episode_index": 4453, "length": 45} +{"episode_index": 4454, "length": 43} +{"episode_index": 4455, "length": 45} +{"episode_index": 4456, "length": 44} +{"episode_index": 4457, "length": 19} +{"episode_index": 4458, "length": 19} +{"episode_index": 4459, "length": 31} +{"episode_index": 4460, "length": 45} +{"episode_index": 4461, "length": 49} +{"episode_index": 4462, "length": 45} +{"episode_index": 4463, "length": 45} +{"episode_index": 4464, "length": 45} +{"episode_index": 4465, "length": 44} +{"episode_index": 4466, "length": 45} +{"episode_index": 4467, "length": 46} +{"episode_index": 4468, "length": 44} +{"episode_index": 4469, "length": 45} +{"episode_index": 4470, "length": 45} +{"episode_index": 4471, "length": 44} +{"episode_index": 4472, "length": 48} +{"episode_index": 4473, "length": 45} +{"episode_index": 4474, "length": 45} +{"episode_index": 4475, "length": 46} +{"episode_index": 4476, "length": 45} +{"episode_index": 4477, "length": 12} +{"episode_index": 4478, "length": 45} +{"episode_index": 4479, "length": 47} +{"episode_index": 4480, "length": 45} +{"episode_index": 4481, "length": 45} +{"episode_index": 4482, "length": 45} +{"episode_index": 4483, "length": 31} +{"episode_index": 4484, "length": 45} +{"episode_index": 4485, "length": 15} +{"episode_index": 4486, "length": 45} +{"episode_index": 4487, "length": 45} +{"episode_index": 4488, "length": 18} +{"episode_index": 4489, "length": 48} +{"episode_index": 4490, "length": 45} +{"episode_index": 4491, "length": 18} +{"episode_index": 4492, "length": 45} +{"episode_index": 4493, "length": 46} +{"episode_index": 4494, "length": 45} +{"episode_index": 4495, "length": 45} +{"episode_index": 4496, "length": 45} +{"episode_index": 4497, "length": 45} +{"episode_index": 4498, "length": 43} +{"episode_index": 4499, "length": 46} +{"episode_index": 4500, "length": 45} +{"episode_index": 4501, "length": 44} +{"episode_index": 4502, "length": 18} +{"episode_index": 4503, "length": 11} +{"episode_index": 4504, "length": 46} +{"episode_index": 4505, "length": 48} +{"episode_index": 4506, "length": 44} +{"episode_index": 4507, "length": 45} +{"episode_index": 4508, "length": 45} +{"episode_index": 4509, "length": 20} +{"episode_index": 4510, "length": 46} +{"episode_index": 4511, "length": 43} +{"episode_index": 4512, "length": 45} +{"episode_index": 4513, "length": 44} +{"episode_index": 4514, "length": 45} +{"episode_index": 4515, "length": 44} +{"episode_index": 4516, "length": 45} +{"episode_index": 4517, "length": 44} +{"episode_index": 4518, "length": 45} +{"episode_index": 4519, "length": 45} +{"episode_index": 4520, "length": 48} +{"episode_index": 4521, "length": 45} +{"episode_index": 4522, "length": 44} +{"episode_index": 4523, "length": 45} +{"episode_index": 4524, "length": 47} +{"episode_index": 4525, "length": 43} +{"episode_index": 4526, "length": 44} +{"episode_index": 4527, "length": 45} +{"episode_index": 4528, "length": 25} +{"episode_index": 4529, "length": 45} +{"episode_index": 4530, "length": 19} +{"episode_index": 4531, "length": 11} +{"episode_index": 4532, "length": 46} +{"episode_index": 4533, "length": 46} +{"episode_index": 4534, "length": 47} +{"episode_index": 4535, "length": 44} +{"episode_index": 4536, "length": 43} +{"episode_index": 4537, "length": 13} +{"episode_index": 4538, "length": 45} +{"episode_index": 4539, "length": 47} +{"episode_index": 4540, "length": 44} +{"episode_index": 4541, "length": 45} +{"episode_index": 4542, "length": 45} +{"episode_index": 4543, "length": 45} +{"episode_index": 4544, "length": 45} +{"episode_index": 4545, "length": 45} +{"episode_index": 4546, "length": 46} +{"episode_index": 4547, "length": 45} +{"episode_index": 4548, "length": 46} +{"episode_index": 4549, "length": 44} +{"episode_index": 4550, "length": 43} +{"episode_index": 4551, "length": 45} +{"episode_index": 4552, "length": 45} +{"episode_index": 4553, "length": 46} +{"episode_index": 4554, "length": 45} +{"episode_index": 4555, "length": 44} +{"episode_index": 4556, "length": 46} +{"episode_index": 4557, "length": 45} +{"episode_index": 4558, "length": 44} +{"episode_index": 4559, "length": 43} +{"episode_index": 4560, "length": 45} +{"episode_index": 4561, "length": 46} +{"episode_index": 4562, "length": 44} +{"episode_index": 4563, "length": 4} +{"episode_index": 4564, "length": 22} +{"episode_index": 4565, "length": 46} +{"episode_index": 4566, "length": 12} +{"episode_index": 4567, "length": 45} +{"episode_index": 4568, "length": 45} +{"episode_index": 4569, "length": 45} +{"episode_index": 4570, "length": 45} +{"episode_index": 4571, "length": 47} +{"episode_index": 4572, "length": 43} +{"episode_index": 4573, "length": 45} +{"episode_index": 4574, "length": 45} +{"episode_index": 4575, "length": 46} +{"episode_index": 4576, "length": 45} +{"episode_index": 4577, "length": 43} +{"episode_index": 4578, "length": 45} +{"episode_index": 4579, "length": 22} +{"episode_index": 4580, "length": 46} +{"episode_index": 4581, "length": 44} +{"episode_index": 4582, "length": 46} +{"episode_index": 4583, "length": 45} +{"episode_index": 4584, "length": 44} +{"episode_index": 4585, "length": 45} +{"episode_index": 4586, "length": 48} +{"episode_index": 4587, "length": 46} +{"episode_index": 4588, "length": 5} +{"episode_index": 4589, "length": 46} +{"episode_index": 4590, "length": 45} +{"episode_index": 4591, "length": 44} +{"episode_index": 4592, "length": 46} +{"episode_index": 4593, "length": 44} +{"episode_index": 4594, "length": 45} +{"episode_index": 4595, "length": 31} +{"episode_index": 4596, "length": 12} +{"episode_index": 4597, "length": 45} +{"episode_index": 4598, "length": 43} +{"episode_index": 4599, "length": 46} +{"episode_index": 4600, "length": 44} +{"episode_index": 4601, "length": 19} +{"episode_index": 4602, "length": 46} +{"episode_index": 4603, "length": 45} +{"episode_index": 4604, "length": 45} +{"episode_index": 4605, "length": 43} +{"episode_index": 4606, "length": 43} +{"episode_index": 4607, "length": 45} +{"episode_index": 4608, "length": 45} +{"episode_index": 4609, "length": 45} +{"episode_index": 4610, "length": 46} +{"episode_index": 4611, "length": 47} +{"episode_index": 4612, "length": 43} +{"episode_index": 4613, "length": 23} +{"episode_index": 4614, "length": 48} +{"episode_index": 4615, "length": 18} +{"episode_index": 4616, "length": 45} +{"episode_index": 4617, "length": 45} +{"episode_index": 4618, "length": 12} +{"episode_index": 4619, "length": 45} +{"episode_index": 4620, "length": 46} +{"episode_index": 4621, "length": 30} +{"episode_index": 4622, "length": 46} +{"episode_index": 4623, "length": 44} +{"episode_index": 4624, "length": 43} +{"episode_index": 4625, "length": 45} +{"episode_index": 4626, "length": 38} +{"episode_index": 4627, "length": 43} +{"episode_index": 4628, "length": 44} +{"episode_index": 4629, "length": 45} +{"episode_index": 4630, "length": 45} +{"episode_index": 4631, "length": 20} +{"episode_index": 4632, "length": 48} +{"episode_index": 4633, "length": 45} +{"episode_index": 4634, "length": 46} +{"episode_index": 4635, "length": 45} +{"episode_index": 4636, "length": 45} +{"episode_index": 4637, "length": 45} +{"episode_index": 4638, "length": 44} +{"episode_index": 4639, "length": 25} +{"episode_index": 4640, "length": 45} +{"episode_index": 4641, "length": 43} +{"episode_index": 4642, "length": 46} +{"episode_index": 4643, "length": 45} +{"episode_index": 4644, "length": 44} +{"episode_index": 4645, "length": 45} +{"episode_index": 4646, "length": 45} +{"episode_index": 4647, "length": 46} +{"episode_index": 4648, "length": 46} +{"episode_index": 4649, "length": 47} +{"episode_index": 4650, "length": 18} +{"episode_index": 4651, "length": 45} +{"episode_index": 4652, "length": 11} +{"episode_index": 4653, "length": 44} +{"episode_index": 4654, "length": 15} +{"episode_index": 4655, "length": 44} +{"episode_index": 4656, "length": 44} +{"episode_index": 4657, "length": 45} +{"episode_index": 4658, "length": 44} +{"episode_index": 4659, "length": 43} +{"episode_index": 4660, "length": 12} +{"episode_index": 4661, "length": 45} +{"episode_index": 4662, "length": 45} +{"episode_index": 4663, "length": 45} +{"episode_index": 4664, "length": 44} +{"episode_index": 4665, "length": 43} +{"episode_index": 4666, "length": 12} +{"episode_index": 4667, "length": 45} +{"episode_index": 4668, "length": 46} +{"episode_index": 4669, "length": 24} +{"episode_index": 4670, "length": 48} +{"episode_index": 4671, "length": 45} +{"episode_index": 4672, "length": 19} +{"episode_index": 4673, "length": 45} +{"episode_index": 4674, "length": 43} +{"episode_index": 4675, "length": 45} +{"episode_index": 4676, "length": 18} +{"episode_index": 4677, "length": 45} +{"episode_index": 4678, "length": 45} +{"episode_index": 4679, "length": 46} +{"episode_index": 4680, "length": 45} +{"episode_index": 4681, "length": 45} +{"episode_index": 4682, "length": 46} +{"episode_index": 4683, "length": 44} +{"episode_index": 4684, "length": 45} +{"episode_index": 4685, "length": 45} +{"episode_index": 4686, "length": 45} +{"episode_index": 4687, "length": 46} +{"episode_index": 4688, "length": 12} +{"episode_index": 4689, "length": 45} +{"episode_index": 4690, "length": 45} +{"episode_index": 4691, "length": 46} +{"episode_index": 4692, "length": 44} +{"episode_index": 4693, "length": 44} +{"episode_index": 4694, "length": 47} +{"episode_index": 4695, "length": 45} +{"episode_index": 4696, "length": 47} +{"episode_index": 4697, "length": 44} +{"episode_index": 4698, "length": 44} +{"episode_index": 4699, "length": 45} +{"episode_index": 4700, "length": 44} +{"episode_index": 4701, "length": 43} +{"episode_index": 4702, "length": 45} +{"episode_index": 4703, "length": 45} +{"episode_index": 4704, "length": 49} +{"episode_index": 4705, "length": 44} +{"episode_index": 4706, "length": 44} +{"episode_index": 4707, "length": 44} +{"episode_index": 4708, "length": 46} +{"episode_index": 4709, "length": 45} +{"episode_index": 4710, "length": 45} +{"episode_index": 4711, "length": 44} +{"episode_index": 4712, "length": 45} +{"episode_index": 4713, "length": 18} +{"episode_index": 4714, "length": 44} +{"episode_index": 4715, "length": 45} +{"episode_index": 4716, "length": 44} +{"episode_index": 4717, "length": 48} +{"episode_index": 4718, "length": 19} +{"episode_index": 4719, "length": 44} +{"episode_index": 4720, "length": 45} +{"episode_index": 4721, "length": 43} +{"episode_index": 4722, "length": 21} +{"episode_index": 4723, "length": 45} +{"episode_index": 4724, "length": 45} +{"episode_index": 4725, "length": 45} +{"episode_index": 4726, "length": 49} +{"episode_index": 4727, "length": 46} +{"episode_index": 4728, "length": 43} +{"episode_index": 4729, "length": 45} +{"episode_index": 4730, "length": 45} +{"episode_index": 4731, "length": 46} +{"episode_index": 4732, "length": 45} +{"episode_index": 4733, "length": 45} +{"episode_index": 4734, "length": 45} +{"episode_index": 4735, "length": 18} +{"episode_index": 4736, "length": 45} +{"episode_index": 4737, "length": 45} +{"episode_index": 4738, "length": 44} +{"episode_index": 4739, "length": 25} +{"episode_index": 4740, "length": 45} +{"episode_index": 4741, "length": 46} +{"episode_index": 4742, "length": 45} +{"episode_index": 4743, "length": 44} +{"episode_index": 4744, "length": 46} +{"episode_index": 4745, "length": 46} +{"episode_index": 4746, "length": 45} +{"episode_index": 4747, "length": 45} +{"episode_index": 4748, "length": 45} +{"episode_index": 4749, "length": 46} +{"episode_index": 4750, "length": 46} +{"episode_index": 4751, "length": 44} +{"episode_index": 4752, "length": 45} +{"episode_index": 4753, "length": 45} +{"episode_index": 4754, "length": 31} +{"episode_index": 4755, "length": 31} +{"episode_index": 4756, "length": 45} +{"episode_index": 4757, "length": 48} +{"episode_index": 4758, "length": 46} +{"episode_index": 4759, "length": 12} +{"episode_index": 4760, "length": 44} +{"episode_index": 4761, "length": 46} +{"episode_index": 4762, "length": 44} +{"episode_index": 4763, "length": 44} +{"episode_index": 4764, "length": 47} +{"episode_index": 4765, "length": 45} +{"episode_index": 4766, "length": 45} +{"episode_index": 4767, "length": 45} +{"episode_index": 4768, "length": 45} +{"episode_index": 4769, "length": 18} +{"episode_index": 4770, "length": 46} +{"episode_index": 4771, "length": 12} +{"episode_index": 4772, "length": 45} +{"episode_index": 4773, "length": 46} +{"episode_index": 4774, "length": 44} +{"episode_index": 4775, "length": 4} +{"episode_index": 4776, "length": 45} +{"episode_index": 4777, "length": 44} +{"episode_index": 4778, "length": 45} +{"episode_index": 4779, "length": 19} +{"episode_index": 4780, "length": 47} +{"episode_index": 4781, "length": 45} +{"episode_index": 4782, "length": 44} +{"episode_index": 4783, "length": 46} +{"episode_index": 4784, "length": 45} +{"episode_index": 4785, "length": 46} +{"episode_index": 4786, "length": 45} +{"episode_index": 4787, "length": 45} +{"episode_index": 4788, "length": 45} +{"episode_index": 4789, "length": 44} +{"episode_index": 4790, "length": 18} +{"episode_index": 4791, "length": 45} +{"episode_index": 4792, "length": 45} +{"episode_index": 4793, "length": 44} +{"episode_index": 4794, "length": 44} +{"episode_index": 4795, "length": 49} +{"episode_index": 4796, "length": 46} +{"episode_index": 4797, "length": 46} +{"episode_index": 4798, "length": 43} +{"episode_index": 4799, "length": 23} +{"episode_index": 4800, "length": 44} +{"episode_index": 4801, "length": 43} +{"episode_index": 4802, "length": 43} +{"episode_index": 4803, "length": 45} +{"episode_index": 4804, "length": 43} +{"episode_index": 4805, "length": 45} +{"episode_index": 4806, "length": 45} +{"episode_index": 4807, "length": 46} +{"episode_index": 4808, "length": 45} +{"episode_index": 4809, "length": 11} +{"episode_index": 4810, "length": 4} +{"episode_index": 4811, "length": 45} +{"episode_index": 4812, "length": 45} +{"episode_index": 4813, "length": 45} +{"episode_index": 4814, "length": 46} +{"episode_index": 4815, "length": 45} +{"episode_index": 4816, "length": 48} +{"episode_index": 4817, "length": 45} +{"episode_index": 4818, "length": 46} +{"episode_index": 4819, "length": 45} +{"episode_index": 4820, "length": 45} +{"episode_index": 4821, "length": 44} +{"episode_index": 4822, "length": 44} +{"episode_index": 4823, "length": 46} +{"episode_index": 4824, "length": 18} +{"episode_index": 4825, "length": 43} +{"episode_index": 4826, "length": 43} +{"episode_index": 4827, "length": 33} +{"episode_index": 4828, "length": 20} +{"episode_index": 4829, "length": 47} +{"episode_index": 4830, "length": 11} +{"episode_index": 4831, "length": 44} +{"episode_index": 4832, "length": 45} +{"episode_index": 4833, "length": 45} +{"episode_index": 4834, "length": 26} +{"episode_index": 4835, "length": 44} +{"episode_index": 4836, "length": 45} +{"episode_index": 4837, "length": 27} +{"episode_index": 4838, "length": 45} +{"episode_index": 4839, "length": 45} +{"episode_index": 4840, "length": 46} +{"episode_index": 4841, "length": 47} +{"episode_index": 4842, "length": 45} +{"episode_index": 4843, "length": 46} +{"episode_index": 4844, "length": 44} +{"episode_index": 4845, "length": 45} +{"episode_index": 4846, "length": 45} +{"episode_index": 4847, "length": 4} +{"episode_index": 4848, "length": 44} +{"episode_index": 4849, "length": 30} +{"episode_index": 4850, "length": 45} +{"episode_index": 4851, "length": 47} +{"episode_index": 4852, "length": 45} +{"episode_index": 4853, "length": 45} +{"episode_index": 4854, "length": 45} +{"episode_index": 4855, "length": 45} +{"episode_index": 4856, "length": 45} +{"episode_index": 4857, "length": 44} +{"episode_index": 4858, "length": 45} +{"episode_index": 4859, "length": 46} +{"episode_index": 4860, "length": 44} +{"episode_index": 4861, "length": 48} +{"episode_index": 4862, "length": 45} +{"episode_index": 4863, "length": 44} +{"episode_index": 4864, "length": 48} +{"episode_index": 4865, "length": 12} +{"episode_index": 4866, "length": 49} +{"episode_index": 4867, "length": 44} +{"episode_index": 4868, "length": 21} +{"episode_index": 4869, "length": 45} +{"episode_index": 4870, "length": 46} +{"episode_index": 4871, "length": 43} +{"episode_index": 4872, "length": 46} +{"episode_index": 4873, "length": 44} +{"episode_index": 4874, "length": 45} +{"episode_index": 4875, "length": 44} +{"episode_index": 4876, "length": 26} +{"episode_index": 4877, "length": 44} +{"episode_index": 4878, "length": 45} +{"episode_index": 4879, "length": 20} +{"episode_index": 4880, "length": 45} +{"episode_index": 4881, "length": 45} +{"episode_index": 4882, "length": 44} +{"episode_index": 4883, "length": 45} +{"episode_index": 4884, "length": 45} +{"episode_index": 4885, "length": 43} +{"episode_index": 4886, "length": 45} +{"episode_index": 4887, "length": 44} +{"episode_index": 4888, "length": 45} +{"episode_index": 4889, "length": 45} +{"episode_index": 4890, "length": 19} +{"episode_index": 4891, "length": 45} +{"episode_index": 4892, "length": 45} +{"episode_index": 4893, "length": 44} +{"episode_index": 4894, "length": 45} +{"episode_index": 4895, "length": 43} +{"episode_index": 4896, "length": 45} +{"episode_index": 4897, "length": 12} +{"episode_index": 4898, "length": 46} +{"episode_index": 4899, "length": 19} +{"episode_index": 4900, "length": 44} +{"episode_index": 4901, "length": 45} +{"episode_index": 4902, "length": 11} +{"episode_index": 4903, "length": 45} +{"episode_index": 4904, "length": 45} +{"episode_index": 4905, "length": 46} +{"episode_index": 4906, "length": 43} +{"episode_index": 4907, "length": 47} +{"episode_index": 4908, "length": 44} +{"episode_index": 4909, "length": 46} +{"episode_index": 4910, "length": 45} +{"episode_index": 4911, "length": 44} +{"episode_index": 4912, "length": 45} +{"episode_index": 4913, "length": 44} +{"episode_index": 4914, "length": 11} +{"episode_index": 4915, "length": 44} +{"episode_index": 4916, "length": 47} +{"episode_index": 4917, "length": 45} +{"episode_index": 4918, "length": 47} +{"episode_index": 4919, "length": 48} +{"episode_index": 4920, "length": 13} +{"episode_index": 4921, "length": 46} +{"episode_index": 4922, "length": 46} +{"episode_index": 4923, "length": 44} +{"episode_index": 4924, "length": 45} +{"episode_index": 4925, "length": 45} +{"episode_index": 4926, "length": 46} +{"episode_index": 4927, "length": 44} +{"episode_index": 4928, "length": 43} +{"episode_index": 4929, "length": 45} +{"episode_index": 4930, "length": 44} +{"episode_index": 4931, "length": 45} +{"episode_index": 4932, "length": 49} +{"episode_index": 4933, "length": 45} +{"episode_index": 4934, "length": 45} +{"episode_index": 4935, "length": 43} +{"episode_index": 4936, "length": 46} +{"episode_index": 4937, "length": 45} +{"episode_index": 4938, "length": 45} +{"episode_index": 4939, "length": 19} +{"episode_index": 4940, "length": 46} +{"episode_index": 4941, "length": 40} +{"episode_index": 4942, "length": 46} +{"episode_index": 4943, "length": 44} +{"episode_index": 4944, "length": 18} +{"episode_index": 4945, "length": 44} +{"episode_index": 4946, "length": 45} +{"episode_index": 4947, "length": 44} +{"episode_index": 4948, "length": 43} +{"episode_index": 4949, "length": 44} +{"episode_index": 4950, "length": 46} +{"episode_index": 4951, "length": 43} +{"episode_index": 4952, "length": 43} +{"episode_index": 4953, "length": 45} +{"episode_index": 4954, "length": 45} +{"episode_index": 4955, "length": 45} +{"episode_index": 4956, "length": 46} +{"episode_index": 4957, "length": 12} +{"episode_index": 4958, "length": 43} +{"episode_index": 4959, "length": 45} +{"episode_index": 4960, "length": 45} +{"episode_index": 4961, "length": 46} +{"episode_index": 4962, "length": 45} +{"episode_index": 4963, "length": 45} +{"episode_index": 4964, "length": 46} +{"episode_index": 4965, "length": 16} +{"episode_index": 4966, "length": 45} +{"episode_index": 4967, "length": 31} +{"episode_index": 4968, "length": 45} +{"episode_index": 4969, "length": 45} +{"episode_index": 4970, "length": 47} +{"episode_index": 4971, "length": 44} +{"episode_index": 4972, "length": 18} +{"episode_index": 4973, "length": 45} +{"episode_index": 4974, "length": 45} +{"episode_index": 4975, "length": 45} +{"episode_index": 4976, "length": 47} +{"episode_index": 4977, "length": 47} +{"episode_index": 4978, "length": 45} +{"episode_index": 4979, "length": 45} +{"episode_index": 4980, "length": 44} +{"episode_index": 4981, "length": 45} +{"episode_index": 4982, "length": 45} +{"episode_index": 4983, "length": 46} +{"episode_index": 4984, "length": 11} +{"episode_index": 4985, "length": 31} +{"episode_index": 4986, "length": 45} +{"episode_index": 4987, "length": 45} +{"episode_index": 4988, "length": 34} +{"episode_index": 4989, "length": 45} +{"episode_index": 4990, "length": 43} +{"episode_index": 4991, "length": 44} +{"episode_index": 4992, "length": 20} +{"episode_index": 4993, "length": 48} +{"episode_index": 4994, "length": 44} +{"episode_index": 4995, "length": 45} +{"episode_index": 4996, "length": 45} +{"episode_index": 4997, "length": 45} +{"episode_index": 4998, "length": 46} +{"episode_index": 4999, "length": 46} +{"episode_index": 5000, "length": 46} +{"episode_index": 5001, "length": 45} +{"episode_index": 5002, "length": 45} +{"episode_index": 5003, "length": 45} +{"episode_index": 5004, "length": 44} +{"episode_index": 5005, "length": 45} +{"episode_index": 5006, "length": 44} +{"episode_index": 5007, "length": 45} +{"episode_index": 5008, "length": 45} +{"episode_index": 5009, "length": 44} +{"episode_index": 5010, "length": 24} +{"episode_index": 5011, "length": 44} +{"episode_index": 5012, "length": 44} +{"episode_index": 5013, "length": 44} +{"episode_index": 5014, "length": 45} +{"episode_index": 5015, "length": 12} +{"episode_index": 5016, "length": 44} +{"episode_index": 5017, "length": 45} +{"episode_index": 5018, "length": 45} +{"episode_index": 5019, "length": 45} +{"episode_index": 5020, "length": 45} +{"episode_index": 5021, "length": 45} +{"episode_index": 5022, "length": 48} +{"episode_index": 5023, "length": 29} +{"episode_index": 5024, "length": 12} +{"episode_index": 5025, "length": 44} +{"episode_index": 5026, "length": 46} +{"episode_index": 5027, "length": 45} +{"episode_index": 5028, "length": 44} +{"episode_index": 5029, "length": 44} +{"episode_index": 5030, "length": 45} +{"episode_index": 5031, "length": 18} +{"episode_index": 5032, "length": 45} +{"episode_index": 5033, "length": 41} +{"episode_index": 5034, "length": 45} +{"episode_index": 5035, "length": 45} +{"episode_index": 5036, "length": 46} +{"episode_index": 5037, "length": 45} +{"episode_index": 5038, "length": 45} +{"episode_index": 5039, "length": 45} +{"episode_index": 5040, "length": 45} +{"episode_index": 5041, "length": 45} +{"episode_index": 5042, "length": 45} +{"episode_index": 5043, "length": 46} +{"episode_index": 5044, "length": 45} +{"episode_index": 5045, "length": 46} +{"episode_index": 5046, "length": 19} +{"episode_index": 5047, "length": 45} +{"episode_index": 5048, "length": 46} +{"episode_index": 5049, "length": 44} +{"episode_index": 5050, "length": 44} +{"episode_index": 5051, "length": 46} +{"episode_index": 5052, "length": 44} +{"episode_index": 5053, "length": 46} +{"episode_index": 5054, "length": 47} +{"episode_index": 5055, "length": 47} +{"episode_index": 5056, "length": 46} +{"episode_index": 5057, "length": 45} +{"episode_index": 5058, "length": 47} +{"episode_index": 5059, "length": 47} +{"episode_index": 5060, "length": 45} +{"episode_index": 5061, "length": 44} +{"episode_index": 5062, "length": 47} +{"episode_index": 5063, "length": 45} +{"episode_index": 5064, "length": 45} +{"episode_index": 5065, "length": 20} +{"episode_index": 5066, "length": 44} +{"episode_index": 5067, "length": 45} +{"episode_index": 5068, "length": 45} +{"episode_index": 5069, "length": 45} +{"episode_index": 5070, "length": 43} +{"episode_index": 5071, "length": 45} +{"episode_index": 5072, "length": 46} +{"episode_index": 5073, "length": 45} +{"episode_index": 5074, "length": 47} +{"episode_index": 5075, "length": 43} +{"episode_index": 5076, "length": 46} +{"episode_index": 5077, "length": 47} +{"episode_index": 5078, "length": 45} +{"episode_index": 5079, "length": 45} +{"episode_index": 5080, "length": 44} +{"episode_index": 5081, "length": 44} +{"episode_index": 5082, "length": 45} +{"episode_index": 5083, "length": 46} +{"episode_index": 5084, "length": 46} +{"episode_index": 5085, "length": 45} +{"episode_index": 5086, "length": 45} +{"episode_index": 5087, "length": 47} +{"episode_index": 5088, "length": 11} +{"episode_index": 5089, "length": 48} +{"episode_index": 5090, "length": 26} +{"episode_index": 5091, "length": 45} +{"episode_index": 5092, "length": 45} +{"episode_index": 5093, "length": 20} +{"episode_index": 5094, "length": 19} +{"episode_index": 5095, "length": 45} +{"episode_index": 5096, "length": 45} +{"episode_index": 5097, "length": 18} +{"episode_index": 5098, "length": 45} +{"episode_index": 5099, "length": 43} +{"episode_index": 5100, "length": 4} +{"episode_index": 5101, "length": 44} +{"episode_index": 5102, "length": 46} +{"episode_index": 5103, "length": 45} +{"episode_index": 5104, "length": 43} +{"episode_index": 5105, "length": 45} +{"episode_index": 5106, "length": 45} +{"episode_index": 5107, "length": 44} +{"episode_index": 5108, "length": 45} +{"episode_index": 5109, "length": 45} +{"episode_index": 5110, "length": 45} +{"episode_index": 5111, "length": 45} +{"episode_index": 5112, "length": 45} +{"episode_index": 5113, "length": 19} +{"episode_index": 5114, "length": 45} +{"episode_index": 5115, "length": 46} +{"episode_index": 5116, "length": 47} +{"episode_index": 5117, "length": 16} +{"episode_index": 5118, "length": 45} +{"episode_index": 5119, "length": 46} +{"episode_index": 5120, "length": 46} +{"episode_index": 5121, "length": 45} +{"episode_index": 5122, "length": 45} +{"episode_index": 5123, "length": 13} +{"episode_index": 5124, "length": 45} +{"episode_index": 5125, "length": 13} +{"episode_index": 5126, "length": 45} +{"episode_index": 5127, "length": 47} +{"episode_index": 5128, "length": 45} +{"episode_index": 5129, "length": 45} +{"episode_index": 5130, "length": 44} +{"episode_index": 5131, "length": 45} +{"episode_index": 5132, "length": 43} +{"episode_index": 5133, "length": 44} +{"episode_index": 5134, "length": 43} +{"episode_index": 5135, "length": 45} +{"episode_index": 5136, "length": 45} +{"episode_index": 5137, "length": 44} +{"episode_index": 5138, "length": 44} +{"episode_index": 5139, "length": 45} +{"episode_index": 5140, "length": 45} +{"episode_index": 5141, "length": 45} +{"episode_index": 5142, "length": 44} +{"episode_index": 5143, "length": 45} +{"episode_index": 5144, "length": 44} +{"episode_index": 5145, "length": 44} +{"episode_index": 5146, "length": 46} +{"episode_index": 5147, "length": 46} +{"episode_index": 5148, "length": 45} +{"episode_index": 5149, "length": 46} +{"episode_index": 5150, "length": 13} +{"episode_index": 5151, "length": 44} +{"episode_index": 5152, "length": 46} +{"episode_index": 5153, "length": 45} +{"episode_index": 5154, "length": 20} +{"episode_index": 5155, "length": 13} +{"episode_index": 5156, "length": 45} +{"episode_index": 5157, "length": 45} +{"episode_index": 5158, "length": 45} +{"episode_index": 5159, "length": 46} +{"episode_index": 5160, "length": 45} +{"episode_index": 5161, "length": 44} +{"episode_index": 5162, "length": 11} +{"episode_index": 5163, "length": 44} +{"episode_index": 5164, "length": 46} +{"episode_index": 5165, "length": 45} +{"episode_index": 5166, "length": 46} +{"episode_index": 5167, "length": 44} +{"episode_index": 5168, "length": 45} +{"episode_index": 5169, "length": 45} +{"episode_index": 5170, "length": 45} +{"episode_index": 5171, "length": 45} +{"episode_index": 5172, "length": 45} +{"episode_index": 5173, "length": 45} +{"episode_index": 5174, "length": 45} +{"episode_index": 5175, "length": 43} +{"episode_index": 5176, "length": 20} +{"episode_index": 5177, "length": 46} +{"episode_index": 5178, "length": 18} +{"episode_index": 5179, "length": 45} +{"episode_index": 5180, "length": 45} +{"episode_index": 5181, "length": 45} +{"episode_index": 5182, "length": 45} +{"episode_index": 5183, "length": 46} +{"episode_index": 5184, "length": 45} +{"episode_index": 5185, "length": 24} +{"episode_index": 5186, "length": 44} +{"episode_index": 5187, "length": 44} +{"episode_index": 5188, "length": 46} +{"episode_index": 5189, "length": 11} +{"episode_index": 5190, "length": 19} +{"episode_index": 5191, "length": 45} +{"episode_index": 5192, "length": 44} +{"episode_index": 5193, "length": 26} +{"episode_index": 5194, "length": 45} +{"episode_index": 5195, "length": 45} +{"episode_index": 5196, "length": 47} +{"episode_index": 5197, "length": 47} +{"episode_index": 5198, "length": 44} +{"episode_index": 5199, "length": 44} +{"episode_index": 5200, "length": 45} +{"episode_index": 5201, "length": 45} +{"episode_index": 5202, "length": 45} +{"episode_index": 5203, "length": 46} +{"episode_index": 5204, "length": 44} +{"episode_index": 5205, "length": 45} +{"episode_index": 5206, "length": 45} +{"episode_index": 5207, "length": 11} +{"episode_index": 5208, "length": 45} +{"episode_index": 5209, "length": 44} +{"episode_index": 5210, "length": 43} +{"episode_index": 5211, "length": 47} +{"episode_index": 5212, "length": 43} +{"episode_index": 5213, "length": 43} +{"episode_index": 5214, "length": 43} +{"episode_index": 5215, "length": 45} +{"episode_index": 5216, "length": 46} +{"episode_index": 5217, "length": 44} +{"episode_index": 5218, "length": 47} +{"episode_index": 5219, "length": 44} +{"episode_index": 5220, "length": 44} +{"episode_index": 5221, "length": 43} +{"episode_index": 5222, "length": 47} +{"episode_index": 5223, "length": 45} +{"episode_index": 5224, "length": 46} +{"episode_index": 5225, "length": 43} +{"episode_index": 5226, "length": 19} +{"episode_index": 5227, "length": 45} +{"episode_index": 5228, "length": 30} +{"episode_index": 5229, "length": 44} +{"episode_index": 5230, "length": 45} +{"episode_index": 5231, "length": 47} +{"episode_index": 5232, "length": 46} +{"episode_index": 5233, "length": 44} +{"episode_index": 5234, "length": 45} +{"episode_index": 5235, "length": 49} +{"episode_index": 5236, "length": 43} +{"episode_index": 5237, "length": 44} +{"episode_index": 5238, "length": 45} +{"episode_index": 5239, "length": 45} +{"episode_index": 5240, "length": 47} +{"episode_index": 5241, "length": 45} +{"episode_index": 5242, "length": 45} +{"episode_index": 5243, "length": 45} +{"episode_index": 5244, "length": 48} +{"episode_index": 5245, "length": 45} +{"episode_index": 5246, "length": 45} +{"episode_index": 5247, "length": 24} +{"episode_index": 5248, "length": 45} +{"episode_index": 5249, "length": 45} +{"episode_index": 5250, "length": 45} +{"episode_index": 5251, "length": 45} +{"episode_index": 5252, "length": 44} +{"episode_index": 5253, "length": 45} +{"episode_index": 5254, "length": 44} +{"episode_index": 5255, "length": 45} +{"episode_index": 5256, "length": 45} +{"episode_index": 5257, "length": 46} +{"episode_index": 5258, "length": 44} +{"episode_index": 5259, "length": 44} +{"episode_index": 5260, "length": 45} +{"episode_index": 5261, "length": 45} +{"episode_index": 5262, "length": 45} +{"episode_index": 5263, "length": 45} +{"episode_index": 5264, "length": 45} +{"episode_index": 5265, "length": 45} +{"episode_index": 5266, "length": 45} +{"episode_index": 5267, "length": 43} +{"episode_index": 5268, "length": 45} +{"episode_index": 5269, "length": 42} +{"episode_index": 5270, "length": 18} +{"episode_index": 5271, "length": 11} +{"episode_index": 5272, "length": 43} +{"episode_index": 5273, "length": 19} +{"episode_index": 5274, "length": 12} +{"episode_index": 5275, "length": 47} +{"episode_index": 5276, "length": 18} +{"episode_index": 5277, "length": 44} +{"episode_index": 5278, "length": 45} +{"episode_index": 5279, "length": 45} +{"episode_index": 5280, "length": 33} +{"episode_index": 5281, "length": 48} +{"episode_index": 5282, "length": 20} +{"episode_index": 5283, "length": 44} +{"episode_index": 5284, "length": 44} +{"episode_index": 5285, "length": 50} +{"episode_index": 5286, "length": 43} +{"episode_index": 5287, "length": 43} +{"episode_index": 5288, "length": 45} +{"episode_index": 5289, "length": 46} +{"episode_index": 5290, "length": 43} +{"episode_index": 5291, "length": 46} +{"episode_index": 5292, "length": 46} +{"episode_index": 5293, "length": 44} +{"episode_index": 5294, "length": 47} +{"episode_index": 5295, "length": 45} +{"episode_index": 5296, "length": 46} +{"episode_index": 5297, "length": 44} +{"episode_index": 5298, "length": 18} +{"episode_index": 5299, "length": 23} +{"episode_index": 5300, "length": 44} +{"episode_index": 5301, "length": 45} +{"episode_index": 5302, "length": 12} +{"episode_index": 5303, "length": 31} +{"episode_index": 5304, "length": 44} +{"episode_index": 5305, "length": 19} +{"episode_index": 5306, "length": 44} +{"episode_index": 5307, "length": 43} +{"episode_index": 5308, "length": 45} +{"episode_index": 5309, "length": 45} +{"episode_index": 5310, "length": 44} +{"episode_index": 5311, "length": 47} +{"episode_index": 5312, "length": 49} +{"episode_index": 5313, "length": 12} +{"episode_index": 5314, "length": 44} +{"episode_index": 5315, "length": 45} +{"episode_index": 5316, "length": 44} +{"episode_index": 5317, "length": 45} +{"episode_index": 5318, "length": 45} +{"episode_index": 5319, "length": 45} +{"episode_index": 5320, "length": 46} +{"episode_index": 5321, "length": 44} +{"episode_index": 5322, "length": 45} +{"episode_index": 5323, "length": 43} +{"episode_index": 5324, "length": 44} +{"episode_index": 5325, "length": 11} +{"episode_index": 5326, "length": 44} +{"episode_index": 5327, "length": 45} +{"episode_index": 5328, "length": 45} +{"episode_index": 5329, "length": 10} +{"episode_index": 5330, "length": 12} +{"episode_index": 5331, "length": 45} +{"episode_index": 5332, "length": 11} +{"episode_index": 5333, "length": 46} +{"episode_index": 5334, "length": 23} +{"episode_index": 5335, "length": 45} +{"episode_index": 5336, "length": 46} +{"episode_index": 5337, "length": 44} +{"episode_index": 5338, "length": 32} +{"episode_index": 5339, "length": 46} +{"episode_index": 5340, "length": 18} +{"episode_index": 5341, "length": 45} +{"episode_index": 5342, "length": 44} +{"episode_index": 5343, "length": 46} +{"episode_index": 5344, "length": 44} +{"episode_index": 5345, "length": 46} +{"episode_index": 5346, "length": 45} +{"episode_index": 5347, "length": 45} +{"episode_index": 5348, "length": 44} +{"episode_index": 5349, "length": 44} +{"episode_index": 5350, "length": 45} +{"episode_index": 5351, "length": 45} +{"episode_index": 5352, "length": 44} +{"episode_index": 5353, "length": 43} +{"episode_index": 5354, "length": 45} +{"episode_index": 5355, "length": 45} +{"episode_index": 5356, "length": 45} +{"episode_index": 5357, "length": 47} +{"episode_index": 5358, "length": 45} +{"episode_index": 5359, "length": 45} +{"episode_index": 5360, "length": 46} +{"episode_index": 5361, "length": 19} +{"episode_index": 5362, "length": 45} +{"episode_index": 5363, "length": 45} +{"episode_index": 5364, "length": 46} +{"episode_index": 5365, "length": 46} +{"episode_index": 5366, "length": 44} +{"episode_index": 5367, "length": 45} +{"episode_index": 5368, "length": 43} +{"episode_index": 5369, "length": 47} +{"episode_index": 5370, "length": 44} +{"episode_index": 5371, "length": 45} +{"episode_index": 5372, "length": 45} +{"episode_index": 5373, "length": 44} +{"episode_index": 5374, "length": 45} +{"episode_index": 5375, "length": 45} +{"episode_index": 5376, "length": 46} +{"episode_index": 5377, "length": 45} +{"episode_index": 5378, "length": 42} +{"episode_index": 5379, "length": 27} +{"episode_index": 5380, "length": 46} +{"episode_index": 5381, "length": 43} +{"episode_index": 5382, "length": 45} +{"episode_index": 5383, "length": 45} +{"episode_index": 5384, "length": 49} +{"episode_index": 5385, "length": 46} +{"episode_index": 5386, "length": 45} +{"episode_index": 5387, "length": 18} +{"episode_index": 5388, "length": 45} +{"episode_index": 5389, "length": 44} +{"episode_index": 5390, "length": 44} +{"episode_index": 5391, "length": 25} +{"episode_index": 5392, "length": 44} +{"episode_index": 5393, "length": 43} +{"episode_index": 5394, "length": 45} +{"episode_index": 5395, "length": 48} +{"episode_index": 5396, "length": 44} +{"episode_index": 5397, "length": 45} +{"episode_index": 5398, "length": 46} +{"episode_index": 5399, "length": 46} +{"episode_index": 5400, "length": 47} +{"episode_index": 5401, "length": 19} +{"episode_index": 5402, "length": 5} +{"episode_index": 5403, "length": 45} +{"episode_index": 5404, "length": 45} +{"episode_index": 5405, "length": 43} +{"episode_index": 5406, "length": 45} +{"episode_index": 5407, "length": 45} +{"episode_index": 5408, "length": 45} +{"episode_index": 5409, "length": 45} +{"episode_index": 5410, "length": 45} +{"episode_index": 5411, "length": 45} +{"episode_index": 5412, "length": 44} +{"episode_index": 5413, "length": 44} +{"episode_index": 5414, "length": 18} +{"episode_index": 5415, "length": 45} +{"episode_index": 5416, "length": 10} +{"episode_index": 5417, "length": 19} +{"episode_index": 5418, "length": 44} +{"episode_index": 5419, "length": 45} +{"episode_index": 5420, "length": 46} +{"episode_index": 5421, "length": 45} +{"episode_index": 5422, "length": 45} +{"episode_index": 5423, "length": 19} +{"episode_index": 5424, "length": 45} +{"episode_index": 5425, "length": 46} +{"episode_index": 5426, "length": 45} +{"episode_index": 5427, "length": 45} +{"episode_index": 5428, "length": 43} +{"episode_index": 5429, "length": 47} +{"episode_index": 5430, "length": 46} +{"episode_index": 5431, "length": 45} +{"episode_index": 5432, "length": 44} +{"episode_index": 5433, "length": 49} +{"episode_index": 5434, "length": 45} +{"episode_index": 5435, "length": 45} +{"episode_index": 5436, "length": 35} +{"episode_index": 5437, "length": 44} +{"episode_index": 5438, "length": 45} +{"episode_index": 5439, "length": 44} +{"episode_index": 5440, "length": 11} +{"episode_index": 5441, "length": 49} +{"episode_index": 5442, "length": 45} +{"episode_index": 5443, "length": 47} +{"episode_index": 5444, "length": 44} +{"episode_index": 5445, "length": 44} +{"episode_index": 5446, "length": 45} +{"episode_index": 5447, "length": 45} +{"episode_index": 5448, "length": 45} +{"episode_index": 5449, "length": 47} +{"episode_index": 5450, "length": 45} +{"episode_index": 5451, "length": 45} +{"episode_index": 5452, "length": 45} +{"episode_index": 5453, "length": 45} +{"episode_index": 5454, "length": 44} +{"episode_index": 5455, "length": 46} +{"episode_index": 5456, "length": 47} +{"episode_index": 5457, "length": 44} +{"episode_index": 5458, "length": 47} +{"episode_index": 5459, "length": 44} +{"episode_index": 5460, "length": 45} +{"episode_index": 5461, "length": 44} +{"episode_index": 5462, "length": 46} +{"episode_index": 5463, "length": 48} +{"episode_index": 5464, "length": 45} +{"episode_index": 5465, "length": 45} +{"episode_index": 5466, "length": 45} +{"episode_index": 5467, "length": 45} +{"episode_index": 5468, "length": 45} +{"episode_index": 5469, "length": 43} +{"episode_index": 5470, "length": 47} +{"episode_index": 5471, "length": 45} +{"episode_index": 5472, "length": 44} +{"episode_index": 5473, "length": 45} +{"episode_index": 5474, "length": 13} +{"episode_index": 5475, "length": 46} +{"episode_index": 5476, "length": 45} +{"episode_index": 5477, "length": 44} +{"episode_index": 5478, "length": 44} +{"episode_index": 5479, "length": 42} +{"episode_index": 5480, "length": 32} +{"episode_index": 5481, "length": 46} +{"episode_index": 5482, "length": 18} +{"episode_index": 5483, "length": 43} +{"episode_index": 5484, "length": 44} +{"episode_index": 5485, "length": 47} +{"episode_index": 5486, "length": 4} +{"episode_index": 5487, "length": 45} +{"episode_index": 5488, "length": 45} +{"episode_index": 5489, "length": 44} +{"episode_index": 5490, "length": 44} +{"episode_index": 5491, "length": 47} +{"episode_index": 5492, "length": 45} +{"episode_index": 5493, "length": 12} +{"episode_index": 5494, "length": 44} +{"episode_index": 5495, "length": 45} +{"episode_index": 5496, "length": 19} +{"episode_index": 5497, "length": 44} +{"episode_index": 5498, "length": 43} +{"episode_index": 5499, "length": 44} +{"episode_index": 5500, "length": 44} +{"episode_index": 5501, "length": 45} +{"episode_index": 5502, "length": 23} +{"episode_index": 5503, "length": 45} +{"episode_index": 5504, "length": 45} +{"episode_index": 5505, "length": 45} +{"episode_index": 5506, "length": 43} +{"episode_index": 5507, "length": 45} +{"episode_index": 5508, "length": 21} +{"episode_index": 5509, "length": 45} +{"episode_index": 5510, "length": 26} +{"episode_index": 5511, "length": 44} +{"episode_index": 5512, "length": 45} +{"episode_index": 5513, "length": 45} +{"episode_index": 5514, "length": 45} +{"episode_index": 5515, "length": 45} +{"episode_index": 5516, "length": 45} +{"episode_index": 5517, "length": 45} +{"episode_index": 5518, "length": 47} +{"episode_index": 5519, "length": 44} +{"episode_index": 5520, "length": 46} +{"episode_index": 5521, "length": 46} +{"episode_index": 5522, "length": 45} +{"episode_index": 5523, "length": 44} +{"episode_index": 5524, "length": 44} +{"episode_index": 5525, "length": 44} +{"episode_index": 5526, "length": 43} +{"episode_index": 5527, "length": 45} +{"episode_index": 5528, "length": 44} +{"episode_index": 5529, "length": 18} +{"episode_index": 5530, "length": 45} +{"episode_index": 5531, "length": 45} +{"episode_index": 5532, "length": 47} +{"episode_index": 5533, "length": 43} +{"episode_index": 5534, "length": 46} +{"episode_index": 5535, "length": 46} +{"episode_index": 5536, "length": 45} +{"episode_index": 5537, "length": 47} +{"episode_index": 5538, "length": 18} +{"episode_index": 5539, "length": 22} +{"episode_index": 5540, "length": 46} +{"episode_index": 5541, "length": 45} +{"episode_index": 5542, "length": 11} +{"episode_index": 5543, "length": 43} +{"episode_index": 5544, "length": 45} +{"episode_index": 5545, "length": 11} +{"episode_index": 5546, "length": 44} +{"episode_index": 5547, "length": 45} +{"episode_index": 5548, "length": 45} +{"episode_index": 5549, "length": 45} +{"episode_index": 5550, "length": 44} +{"episode_index": 5551, "length": 46} +{"episode_index": 5552, "length": 47} +{"episode_index": 5553, "length": 47} +{"episode_index": 5554, "length": 44} +{"episode_index": 5555, "length": 28} +{"episode_index": 5556, "length": 45} +{"episode_index": 5557, "length": 43} +{"episode_index": 5558, "length": 44} +{"episode_index": 5559, "length": 44} +{"episode_index": 5560, "length": 43} +{"episode_index": 5561, "length": 46} +{"episode_index": 5562, "length": 44} +{"episode_index": 5563, "length": 33} +{"episode_index": 5564, "length": 24} +{"episode_index": 5565, "length": 46} +{"episode_index": 5566, "length": 43} +{"episode_index": 5567, "length": 45} +{"episode_index": 5568, "length": 45} +{"episode_index": 5569, "length": 44} +{"episode_index": 5570, "length": 45} +{"episode_index": 5571, "length": 45} +{"episode_index": 5572, "length": 45} +{"episode_index": 5573, "length": 43} +{"episode_index": 5574, "length": 45} +{"episode_index": 5575, "length": 19} +{"episode_index": 5576, "length": 47} +{"episode_index": 5577, "length": 45} +{"episode_index": 5578, "length": 45} +{"episode_index": 5579, "length": 45} +{"episode_index": 5580, "length": 46} +{"episode_index": 5581, "length": 12} +{"episode_index": 5582, "length": 46} +{"episode_index": 5583, "length": 47} +{"episode_index": 5584, "length": 47} +{"episode_index": 5585, "length": 45} +{"episode_index": 5586, "length": 45} +{"episode_index": 5587, "length": 45} +{"episode_index": 5588, "length": 30} +{"episode_index": 5589, "length": 45} +{"episode_index": 5590, "length": 47} +{"episode_index": 5591, "length": 45} +{"episode_index": 5592, "length": 45} +{"episode_index": 5593, "length": 45} +{"episode_index": 5594, "length": 45} +{"episode_index": 5595, "length": 46} +{"episode_index": 5596, "length": 45} +{"episode_index": 5597, "length": 45} +{"episode_index": 5598, "length": 12} +{"episode_index": 5599, "length": 26} +{"episode_index": 5600, "length": 44} +{"episode_index": 5601, "length": 19} +{"episode_index": 5602, "length": 44} +{"episode_index": 5603, "length": 45} +{"episode_index": 5604, "length": 26} +{"episode_index": 5605, "length": 46} +{"episode_index": 5606, "length": 45} +{"episode_index": 5607, "length": 46} +{"episode_index": 5608, "length": 44} +{"episode_index": 5609, "length": 45} +{"episode_index": 5610, "length": 44} +{"episode_index": 5611, "length": 47} +{"episode_index": 5612, "length": 44} +{"episode_index": 5613, "length": 47} +{"episode_index": 5614, "length": 46} +{"episode_index": 5615, "length": 19} +{"episode_index": 5616, "length": 46} +{"episode_index": 5617, "length": 45} +{"episode_index": 5618, "length": 43} +{"episode_index": 5619, "length": 45} +{"episode_index": 5620, "length": 47} +{"episode_index": 5621, "length": 46} +{"episode_index": 5622, "length": 46} +{"episode_index": 5623, "length": 44} +{"episode_index": 5624, "length": 45} +{"episode_index": 5625, "length": 45} +{"episode_index": 5626, "length": 45} +{"episode_index": 5627, "length": 45} +{"episode_index": 5628, "length": 45} +{"episode_index": 5629, "length": 45} +{"episode_index": 5630, "length": 19} +{"episode_index": 5631, "length": 48} +{"episode_index": 5632, "length": 44} +{"episode_index": 5633, "length": 44} +{"episode_index": 5634, "length": 44} +{"episode_index": 5635, "length": 46} +{"episode_index": 5636, "length": 45} +{"episode_index": 5637, "length": 45} +{"episode_index": 5638, "length": 44} +{"episode_index": 5639, "length": 44} +{"episode_index": 5640, "length": 45} +{"episode_index": 5641, "length": 44} +{"episode_index": 5642, "length": 45} +{"episode_index": 5643, "length": 44} +{"episode_index": 5644, "length": 45} +{"episode_index": 5645, "length": 44} +{"episode_index": 5646, "length": 43} +{"episode_index": 5647, "length": 45} +{"episode_index": 5648, "length": 11} +{"episode_index": 5649, "length": 44} +{"episode_index": 5650, "length": 46} +{"episode_index": 5651, "length": 45} +{"episode_index": 5652, "length": 46} +{"episode_index": 5653, "length": 44} +{"episode_index": 5654, "length": 45} +{"episode_index": 5655, "length": 47} +{"episode_index": 5656, "length": 44} +{"episode_index": 5657, "length": 46} +{"episode_index": 5658, "length": 48} +{"episode_index": 5659, "length": 45} +{"episode_index": 5660, "length": 44} +{"episode_index": 5661, "length": 45} +{"episode_index": 5662, "length": 46} +{"episode_index": 5663, "length": 11} +{"episode_index": 5664, "length": 11} +{"episode_index": 5665, "length": 45} +{"episode_index": 5666, "length": 44} +{"episode_index": 5667, "length": 24} +{"episode_index": 5668, "length": 46} +{"episode_index": 5669, "length": 47} +{"episode_index": 5670, "length": 45} +{"episode_index": 5671, "length": 47} +{"episode_index": 5672, "length": 20} +{"episode_index": 5673, "length": 44} +{"episode_index": 5674, "length": 45} +{"episode_index": 5675, "length": 47} +{"episode_index": 5676, "length": 19} +{"episode_index": 5677, "length": 44} +{"episode_index": 5678, "length": 45} +{"episode_index": 5679, "length": 44} +{"episode_index": 5680, "length": 47} +{"episode_index": 5681, "length": 45} +{"episode_index": 5682, "length": 45} +{"episode_index": 5683, "length": 44} +{"episode_index": 5684, "length": 26} +{"episode_index": 5685, "length": 44} +{"episode_index": 5686, "length": 44} +{"episode_index": 5687, "length": 46} +{"episode_index": 5688, "length": 44} +{"episode_index": 5689, "length": 11} +{"episode_index": 5690, "length": 45} +{"episode_index": 5691, "length": 46} +{"episode_index": 5692, "length": 44} +{"episode_index": 5693, "length": 10} +{"episode_index": 5694, "length": 48} +{"episode_index": 5695, "length": 45} +{"episode_index": 5696, "length": 45} +{"episode_index": 5697, "length": 45} +{"episode_index": 5698, "length": 43} +{"episode_index": 5699, "length": 44} +{"episode_index": 5700, "length": 4} +{"episode_index": 5701, "length": 45} +{"episode_index": 5702, "length": 18} +{"episode_index": 5703, "length": 27} +{"episode_index": 5704, "length": 44} +{"episode_index": 5705, "length": 45} +{"episode_index": 5706, "length": 44} +{"episode_index": 5707, "length": 47} +{"episode_index": 5708, "length": 12} +{"episode_index": 5709, "length": 45} +{"episode_index": 5710, "length": 44} +{"episode_index": 5711, "length": 45} +{"episode_index": 5712, "length": 46} +{"episode_index": 5713, "length": 33} +{"episode_index": 5714, "length": 45} +{"episode_index": 5715, "length": 44} +{"episode_index": 5716, "length": 45} +{"episode_index": 5717, "length": 46} +{"episode_index": 5718, "length": 45} +{"episode_index": 5719, "length": 45} +{"episode_index": 5720, "length": 45} +{"episode_index": 5721, "length": 44} +{"episode_index": 5722, "length": 45} +{"episode_index": 5723, "length": 45} +{"episode_index": 5724, "length": 18} +{"episode_index": 5725, "length": 44} +{"episode_index": 5726, "length": 44} +{"episode_index": 5727, "length": 20} +{"episode_index": 5728, "length": 45} +{"episode_index": 5729, "length": 44} +{"episode_index": 5730, "length": 19} +{"episode_index": 5731, "length": 25} +{"episode_index": 5732, "length": 46} +{"episode_index": 5733, "length": 24} +{"episode_index": 5734, "length": 43} +{"episode_index": 5735, "length": 46} +{"episode_index": 5736, "length": 44} +{"episode_index": 5737, "length": 11} +{"episode_index": 5738, "length": 44} +{"episode_index": 5739, "length": 46} +{"episode_index": 5740, "length": 45} +{"episode_index": 5741, "length": 49} +{"episode_index": 5742, "length": 45} +{"episode_index": 5743, "length": 45} +{"episode_index": 5744, "length": 45} +{"episode_index": 5745, "length": 45} +{"episode_index": 5746, "length": 38} +{"episode_index": 5747, "length": 19} +{"episode_index": 5748, "length": 43} +{"episode_index": 5749, "length": 44} +{"episode_index": 5750, "length": 19} +{"episode_index": 5751, "length": 45} +{"episode_index": 5752, "length": 45} +{"episode_index": 5753, "length": 45} +{"episode_index": 5754, "length": 45} +{"episode_index": 5755, "length": 45} +{"episode_index": 5756, "length": 11} +{"episode_index": 5757, "length": 45} +{"episode_index": 5758, "length": 18} +{"episode_index": 5759, "length": 45} +{"episode_index": 5760, "length": 43} +{"episode_index": 5761, "length": 45} +{"episode_index": 5762, "length": 44} +{"episode_index": 5763, "length": 45} +{"episode_index": 5764, "length": 44} +{"episode_index": 5765, "length": 21} +{"episode_index": 5766, "length": 44} +{"episode_index": 5767, "length": 44} +{"episode_index": 5768, "length": 47} +{"episode_index": 5769, "length": 45} +{"episode_index": 5770, "length": 45} +{"episode_index": 5771, "length": 45} +{"episode_index": 5772, "length": 45} +{"episode_index": 5773, "length": 19} +{"episode_index": 5774, "length": 49} +{"episode_index": 5775, "length": 44} +{"episode_index": 5776, "length": 44} +{"episode_index": 5777, "length": 45} +{"episode_index": 5778, "length": 45} +{"episode_index": 5779, "length": 43} +{"episode_index": 5780, "length": 45} +{"episode_index": 5781, "length": 43} +{"episode_index": 5782, "length": 43} +{"episode_index": 5783, "length": 45} +{"episode_index": 5784, "length": 45} +{"episode_index": 5785, "length": 44} +{"episode_index": 5786, "length": 5} +{"episode_index": 5787, "length": 45} +{"episode_index": 5788, "length": 45} +{"episode_index": 5789, "length": 46} +{"episode_index": 5790, "length": 46} +{"episode_index": 5791, "length": 45} +{"episode_index": 5792, "length": 44} +{"episode_index": 5793, "length": 44} +{"episode_index": 5794, "length": 43} +{"episode_index": 5795, "length": 46} +{"episode_index": 5796, "length": 45} +{"episode_index": 5797, "length": 44} +{"episode_index": 5798, "length": 45} +{"episode_index": 5799, "length": 12} +{"episode_index": 5800, "length": 44} +{"episode_index": 5801, "length": 45} +{"episode_index": 5802, "length": 13} +{"episode_index": 5803, "length": 45} +{"episode_index": 5804, "length": 45} +{"episode_index": 5805, "length": 30} +{"episode_index": 5806, "length": 44} +{"episode_index": 5807, "length": 45} +{"episode_index": 5808, "length": 45} +{"episode_index": 5809, "length": 30} +{"episode_index": 5810, "length": 46} +{"episode_index": 5811, "length": 18} +{"episode_index": 5812, "length": 45} +{"episode_index": 5813, "length": 45} +{"episode_index": 5814, "length": 46} +{"episode_index": 5815, "length": 19} +{"episode_index": 5816, "length": 45} +{"episode_index": 5817, "length": 24} +{"episode_index": 5818, "length": 45} +{"episode_index": 5819, "length": 45} +{"episode_index": 5820, "length": 44} +{"episode_index": 5821, "length": 47} +{"episode_index": 5822, "length": 43} +{"episode_index": 5823, "length": 43} +{"episode_index": 5824, "length": 43} +{"episode_index": 5825, "length": 10} +{"episode_index": 5826, "length": 47} +{"episode_index": 5827, "length": 45} +{"episode_index": 5828, "length": 45} +{"episode_index": 5829, "length": 44} +{"episode_index": 5830, "length": 45} +{"episode_index": 5831, "length": 44} +{"episode_index": 5832, "length": 45} +{"episode_index": 5833, "length": 34} +{"episode_index": 5834, "length": 47} +{"episode_index": 5835, "length": 46} +{"episode_index": 5836, "length": 43} +{"episode_index": 5837, "length": 45} +{"episode_index": 5838, "length": 43} +{"episode_index": 5839, "length": 45} +{"episode_index": 5840, "length": 45} +{"episode_index": 5841, "length": 48} +{"episode_index": 5842, "length": 47} +{"episode_index": 5843, "length": 46} +{"episode_index": 5844, "length": 44} +{"episode_index": 5845, "length": 45} +{"episode_index": 5846, "length": 45} +{"episode_index": 5847, "length": 44} +{"episode_index": 5848, "length": 46} +{"episode_index": 5849, "length": 45} +{"episode_index": 5850, "length": 45} +{"episode_index": 5851, "length": 45} +{"episode_index": 5852, "length": 44} +{"episode_index": 5853, "length": 45} +{"episode_index": 5854, "length": 45} +{"episode_index": 5855, "length": 47} +{"episode_index": 5856, "length": 44} +{"episode_index": 5857, "length": 47} +{"episode_index": 5858, "length": 24} +{"episode_index": 5859, "length": 43} +{"episode_index": 5860, "length": 46} +{"episode_index": 5861, "length": 19} +{"episode_index": 5862, "length": 44} +{"episode_index": 5863, "length": 48} +{"episode_index": 5864, "length": 45} +{"episode_index": 5865, "length": 45} +{"episode_index": 5866, "length": 46} +{"episode_index": 5867, "length": 44} +{"episode_index": 5868, "length": 43} +{"episode_index": 5869, "length": 46} +{"episode_index": 5870, "length": 45} +{"episode_index": 5871, "length": 48} +{"episode_index": 5872, "length": 26} +{"episode_index": 5873, "length": 11} +{"episode_index": 5874, "length": 45} +{"episode_index": 5875, "length": 47} +{"episode_index": 5876, "length": 45} +{"episode_index": 5877, "length": 44} +{"episode_index": 5878, "length": 46} +{"episode_index": 5879, "length": 45} +{"episode_index": 5880, "length": 45} +{"episode_index": 5881, "length": 45} +{"episode_index": 5882, "length": 22} +{"episode_index": 5883, "length": 30} +{"episode_index": 5884, "length": 44} +{"episode_index": 5885, "length": 31} +{"episode_index": 5886, "length": 44} +{"episode_index": 5887, "length": 45} +{"episode_index": 5888, "length": 45} +{"episode_index": 5889, "length": 46} +{"episode_index": 5890, "length": 45} +{"episode_index": 5891, "length": 45} +{"episode_index": 5892, "length": 13} +{"episode_index": 5893, "length": 45} +{"episode_index": 5894, "length": 45} +{"episode_index": 5895, "length": 45} +{"episode_index": 5896, "length": 44} +{"episode_index": 5897, "length": 43} +{"episode_index": 5898, "length": 45} +{"episode_index": 5899, "length": 43} +{"episode_index": 5900, "length": 19} +{"episode_index": 5901, "length": 44} +{"episode_index": 5902, "length": 43} +{"episode_index": 5903, "length": 44} +{"episode_index": 5904, "length": 45} +{"episode_index": 5905, "length": 20} +{"episode_index": 5906, "length": 44} +{"episode_index": 5907, "length": 49} +{"episode_index": 5908, "length": 43} +{"episode_index": 5909, "length": 46} +{"episode_index": 5910, "length": 27} +{"episode_index": 5911, "length": 47} +{"episode_index": 5912, "length": 47} +{"episode_index": 5913, "length": 11} +{"episode_index": 5914, "length": 44} +{"episode_index": 5915, "length": 45} +{"episode_index": 5916, "length": 45} +{"episode_index": 5917, "length": 45} +{"episode_index": 5918, "length": 46} +{"episode_index": 5919, "length": 45} +{"episode_index": 5920, "length": 47} +{"episode_index": 5921, "length": 45} +{"episode_index": 5922, "length": 45} +{"episode_index": 5923, "length": 45} +{"episode_index": 5924, "length": 43} +{"episode_index": 5925, "length": 45} +{"episode_index": 5926, "length": 45} +{"episode_index": 5927, "length": 43} +{"episode_index": 5928, "length": 43} +{"episode_index": 5929, "length": 44} +{"episode_index": 5930, "length": 46} +{"episode_index": 5931, "length": 46} +{"episode_index": 5932, "length": 46} +{"episode_index": 5933, "length": 45} +{"episode_index": 5934, "length": 44} +{"episode_index": 5935, "length": 45} +{"episode_index": 5936, "length": 46} +{"episode_index": 5937, "length": 44} +{"episode_index": 5938, "length": 44} +{"episode_index": 5939, "length": 47} +{"episode_index": 5940, "length": 45} +{"episode_index": 5941, "length": 45} +{"episode_index": 5942, "length": 43} +{"episode_index": 5943, "length": 47} +{"episode_index": 5944, "length": 49} +{"episode_index": 5945, "length": 44} +{"episode_index": 5946, "length": 48} +{"episode_index": 5947, "length": 45} +{"episode_index": 5948, "length": 44} +{"episode_index": 5949, "length": 46} +{"episode_index": 5950, "length": 45} +{"episode_index": 5951, "length": 43} +{"episode_index": 5952, "length": 47} +{"episode_index": 5953, "length": 45} +{"episode_index": 5954, "length": 19} +{"episode_index": 5955, "length": 46} +{"episode_index": 5956, "length": 43} +{"episode_index": 5957, "length": 46} +{"episode_index": 5958, "length": 46} +{"episode_index": 5959, "length": 20} +{"episode_index": 5960, "length": 43} +{"episode_index": 5961, "length": 48} +{"episode_index": 5962, "length": 46} +{"episode_index": 5963, "length": 32} +{"episode_index": 5964, "length": 45} +{"episode_index": 5965, "length": 45} +{"episode_index": 5966, "length": 44} +{"episode_index": 5967, "length": 44} +{"episode_index": 5968, "length": 45} +{"episode_index": 5969, "length": 45} +{"episode_index": 5970, "length": 47} +{"episode_index": 5971, "length": 46} +{"episode_index": 5972, "length": 46} +{"episode_index": 5973, "length": 11} +{"episode_index": 5974, "length": 48} +{"episode_index": 5975, "length": 46} +{"episode_index": 5976, "length": 47} +{"episode_index": 5977, "length": 44} +{"episode_index": 5978, "length": 26} +{"episode_index": 5979, "length": 18} +{"episode_index": 5980, "length": 45} +{"episode_index": 5981, "length": 45} +{"episode_index": 5982, "length": 45} +{"episode_index": 5983, "length": 45} +{"episode_index": 5984, "length": 45} +{"episode_index": 5985, "length": 44} +{"episode_index": 5986, "length": 43} +{"episode_index": 5987, "length": 48} +{"episode_index": 5988, "length": 42} +{"episode_index": 5989, "length": 44} +{"episode_index": 5990, "length": 45} +{"episode_index": 5991, "length": 44} +{"episode_index": 5992, "length": 45} +{"episode_index": 5993, "length": 45} +{"episode_index": 5994, "length": 48} +{"episode_index": 5995, "length": 40} +{"episode_index": 5996, "length": 19} +{"episode_index": 5997, "length": 45} +{"episode_index": 5998, "length": 46} +{"episode_index": 5999, "length": 43} +{"episode_index": 6000, "length": 44} +{"episode_index": 6001, "length": 47} +{"episode_index": 6002, "length": 18} +{"episode_index": 6003, "length": 45} +{"episode_index": 6004, "length": 45} +{"episode_index": 6005, "length": 46} +{"episode_index": 6006, "length": 44} +{"episode_index": 6007, "length": 45} +{"episode_index": 6008, "length": 45} +{"episode_index": 6009, "length": 45} +{"episode_index": 6010, "length": 47} +{"episode_index": 6011, "length": 44} +{"episode_index": 6012, "length": 21} +{"episode_index": 6013, "length": 44} +{"episode_index": 6014, "length": 10} +{"episode_index": 6015, "length": 44} +{"episode_index": 6016, "length": 45} +{"episode_index": 6017, "length": 45} +{"episode_index": 6018, "length": 45} +{"episode_index": 6019, "length": 45} +{"episode_index": 6020, "length": 45} +{"episode_index": 6021, "length": 48} +{"episode_index": 6022, "length": 44} +{"episode_index": 6023, "length": 47} +{"episode_index": 6024, "length": 18} +{"episode_index": 6025, "length": 45} +{"episode_index": 6026, "length": 45} +{"episode_index": 6027, "length": 44} +{"episode_index": 6028, "length": 19} +{"episode_index": 6029, "length": 45} +{"episode_index": 6030, "length": 46} +{"episode_index": 6031, "length": 43} +{"episode_index": 6032, "length": 45} +{"episode_index": 6033, "length": 47} +{"episode_index": 6034, "length": 45} +{"episode_index": 6035, "length": 45} +{"episode_index": 6036, "length": 45} +{"episode_index": 6037, "length": 49} +{"episode_index": 6038, "length": 45} +{"episode_index": 6039, "length": 44} +{"episode_index": 6040, "length": 25} +{"episode_index": 6041, "length": 29} +{"episode_index": 6042, "length": 47} +{"episode_index": 6043, "length": 44} +{"episode_index": 6044, "length": 19} +{"episode_index": 6045, "length": 45} +{"episode_index": 6046, "length": 45} +{"episode_index": 6047, "length": 11} +{"episode_index": 6048, "length": 20} +{"episode_index": 6049, "length": 18} +{"episode_index": 6050, "length": 45} +{"episode_index": 6051, "length": 44} +{"episode_index": 6052, "length": 49} +{"episode_index": 6053, "length": 43} +{"episode_index": 6054, "length": 47} +{"episode_index": 6055, "length": 45} +{"episode_index": 6056, "length": 45} +{"episode_index": 6057, "length": 44} +{"episode_index": 6058, "length": 43} +{"episode_index": 6059, "length": 45} +{"episode_index": 6060, "length": 45} +{"episode_index": 6061, "length": 19} +{"episode_index": 6062, "length": 45} +{"episode_index": 6063, "length": 44} +{"episode_index": 6064, "length": 45} +{"episode_index": 6065, "length": 45} +{"episode_index": 6066, "length": 44} +{"episode_index": 6067, "length": 43} +{"episode_index": 6068, "length": 48} +{"episode_index": 6069, "length": 44} +{"episode_index": 6070, "length": 45} +{"episode_index": 6071, "length": 45} +{"episode_index": 6072, "length": 46} +{"episode_index": 6073, "length": 45} +{"episode_index": 6074, "length": 46} +{"episode_index": 6075, "length": 44} +{"episode_index": 6076, "length": 47} +{"episode_index": 6077, "length": 45} +{"episode_index": 6078, "length": 44} +{"episode_index": 6079, "length": 47} +{"episode_index": 6080, "length": 45} +{"episode_index": 6081, "length": 45} +{"episode_index": 6082, "length": 43} +{"episode_index": 6083, "length": 43} +{"episode_index": 6084, "length": 45} +{"episode_index": 6085, "length": 43} +{"episode_index": 6086, "length": 45} +{"episode_index": 6087, "length": 45} +{"episode_index": 6088, "length": 11} +{"episode_index": 6089, "length": 43} +{"episode_index": 6090, "length": 45} +{"episode_index": 6091, "length": 45} +{"episode_index": 6092, "length": 18} +{"episode_index": 6093, "length": 49} +{"episode_index": 6094, "length": 48} +{"episode_index": 6095, "length": 46} +{"episode_index": 6096, "length": 19} +{"episode_index": 6097, "length": 45} +{"episode_index": 6098, "length": 5} +{"episode_index": 6099, "length": 46} +{"episode_index": 6100, "length": 45} +{"episode_index": 6101, "length": 47} +{"episode_index": 6102, "length": 44} +{"episode_index": 6103, "length": 44} +{"episode_index": 6104, "length": 45} +{"episode_index": 6105, "length": 44} +{"episode_index": 6106, "length": 44} +{"episode_index": 6107, "length": 43} +{"episode_index": 6108, "length": 45} +{"episode_index": 6109, "length": 45} +{"episode_index": 6110, "length": 45} +{"episode_index": 6111, "length": 43} +{"episode_index": 6112, "length": 44} +{"episode_index": 6113, "length": 45} +{"episode_index": 6114, "length": 45} +{"episode_index": 6115, "length": 31} +{"episode_index": 6116, "length": 45} +{"episode_index": 6117, "length": 48} +{"episode_index": 6118, "length": 44} +{"episode_index": 6119, "length": 45} +{"episode_index": 6120, "length": 45} +{"episode_index": 6121, "length": 45} +{"episode_index": 6122, "length": 29} +{"episode_index": 6123, "length": 46} +{"episode_index": 6124, "length": 46} +{"episode_index": 6125, "length": 18} +{"episode_index": 6126, "length": 44} +{"episode_index": 6127, "length": 45} +{"episode_index": 6128, "length": 44} +{"episode_index": 6129, "length": 45} +{"episode_index": 6130, "length": 46} +{"episode_index": 6131, "length": 46} +{"episode_index": 6132, "length": 43} +{"episode_index": 6133, "length": 44} +{"episode_index": 6134, "length": 45} +{"episode_index": 6135, "length": 45} +{"episode_index": 6136, "length": 47} +{"episode_index": 6137, "length": 43} +{"episode_index": 6138, "length": 44} +{"episode_index": 6139, "length": 45} +{"episode_index": 6140, "length": 46} +{"episode_index": 6141, "length": 45} +{"episode_index": 6142, "length": 47} +{"episode_index": 6143, "length": 45} +{"episode_index": 6144, "length": 45} +{"episode_index": 6145, "length": 46} +{"episode_index": 6146, "length": 45} +{"episode_index": 6147, "length": 22} +{"episode_index": 6148, "length": 46} +{"episode_index": 6149, "length": 45} +{"episode_index": 6150, "length": 46} +{"episode_index": 6151, "length": 44} +{"episode_index": 6152, "length": 43} +{"episode_index": 6153, "length": 11} +{"episode_index": 6154, "length": 45} +{"episode_index": 6155, "length": 47} +{"episode_index": 6156, "length": 45} +{"episode_index": 6157, "length": 32} +{"episode_index": 6158, "length": 45} +{"episode_index": 6159, "length": 45} +{"episode_index": 6160, "length": 19} +{"episode_index": 6161, "length": 45} +{"episode_index": 6162, "length": 45} +{"episode_index": 6163, "length": 20} +{"episode_index": 6164, "length": 43} +{"episode_index": 6165, "length": 45} +{"episode_index": 6166, "length": 45} +{"episode_index": 6167, "length": 45} +{"episode_index": 6168, "length": 45} +{"episode_index": 6169, "length": 46} +{"episode_index": 6170, "length": 44} +{"episode_index": 6171, "length": 44} +{"episode_index": 6172, "length": 46} +{"episode_index": 6173, "length": 47} +{"episode_index": 6174, "length": 47} +{"episode_index": 6175, "length": 44} +{"episode_index": 6176, "length": 44} +{"episode_index": 6177, "length": 44} +{"episode_index": 6178, "length": 44} +{"episode_index": 6179, "length": 45} +{"episode_index": 6180, "length": 44} +{"episode_index": 6181, "length": 46} +{"episode_index": 6182, "length": 46} +{"episode_index": 6183, "length": 47} +{"episode_index": 6184, "length": 19} +{"episode_index": 6185, "length": 45} +{"episode_index": 6186, "length": 45} +{"episode_index": 6187, "length": 45} +{"episode_index": 6188, "length": 45} +{"episode_index": 6189, "length": 45} +{"episode_index": 6190, "length": 45} +{"episode_index": 6191, "length": 47} +{"episode_index": 6192, "length": 45} +{"episode_index": 6193, "length": 45} +{"episode_index": 6194, "length": 44} +{"episode_index": 6195, "length": 47} +{"episode_index": 6196, "length": 45} +{"episode_index": 6197, "length": 47} +{"episode_index": 6198, "length": 19} +{"episode_index": 6199, "length": 44} +{"episode_index": 6200, "length": 46} +{"episode_index": 6201, "length": 45} +{"episode_index": 6202, "length": 44} +{"episode_index": 6203, "length": 18} +{"episode_index": 6204, "length": 45} +{"episode_index": 6205, "length": 46} +{"episode_index": 6206, "length": 43} +{"episode_index": 6207, "length": 45} +{"episode_index": 6208, "length": 11} +{"episode_index": 6209, "length": 45} +{"episode_index": 6210, "length": 45} +{"episode_index": 6211, "length": 45} +{"episode_index": 6212, "length": 45} +{"episode_index": 6213, "length": 24} +{"episode_index": 6214, "length": 45} +{"episode_index": 6215, "length": 45} +{"episode_index": 6216, "length": 47} +{"episode_index": 6217, "length": 45} +{"episode_index": 6218, "length": 45} +{"episode_index": 6219, "length": 44} +{"episode_index": 6220, "length": 45} +{"episode_index": 6221, "length": 11} +{"episode_index": 6222, "length": 45} +{"episode_index": 6223, "length": 10} +{"episode_index": 6224, "length": 45} +{"episode_index": 6225, "length": 45} +{"episode_index": 6226, "length": 45} +{"episode_index": 6227, "length": 18} +{"episode_index": 6228, "length": 45} +{"episode_index": 6229, "length": 47} +{"episode_index": 6230, "length": 46} +{"episode_index": 6231, "length": 45} +{"episode_index": 6232, "length": 45} +{"episode_index": 6233, "length": 45} +{"episode_index": 6234, "length": 45} +{"episode_index": 6235, "length": 45} +{"episode_index": 6236, "length": 43} +{"episode_index": 6237, "length": 18} +{"episode_index": 6238, "length": 45} +{"episode_index": 6239, "length": 43} +{"episode_index": 6240, "length": 46} +{"episode_index": 6241, "length": 24} +{"episode_index": 6242, "length": 43} +{"episode_index": 6243, "length": 45} +{"episode_index": 6244, "length": 43} +{"episode_index": 6245, "length": 45} +{"episode_index": 6246, "length": 47} +{"episode_index": 6247, "length": 47} +{"episode_index": 6248, "length": 48} +{"episode_index": 6249, "length": 48} +{"episode_index": 6250, "length": 44} +{"episode_index": 6251, "length": 47} +{"episode_index": 6252, "length": 44} +{"episode_index": 6253, "length": 43} +{"episode_index": 6254, "length": 46} +{"episode_index": 6255, "length": 18} +{"episode_index": 6256, "length": 44} +{"episode_index": 6257, "length": 46} +{"episode_index": 6258, "length": 47} +{"episode_index": 6259, "length": 43} +{"episode_index": 6260, "length": 46} +{"episode_index": 6261, "length": 46} +{"episode_index": 6262, "length": 44} +{"episode_index": 6263, "length": 45} +{"episode_index": 6264, "length": 41} +{"episode_index": 6265, "length": 45} +{"episode_index": 6266, "length": 44} +{"episode_index": 6267, "length": 45} +{"episode_index": 6268, "length": 45} +{"episode_index": 6269, "length": 44} +{"episode_index": 6270, "length": 45} +{"episode_index": 6271, "length": 45} +{"episode_index": 6272, "length": 46} +{"episode_index": 6273, "length": 46} +{"episode_index": 6274, "length": 45} +{"episode_index": 6275, "length": 46} +{"episode_index": 6276, "length": 44} +{"episode_index": 6277, "length": 46} +{"episode_index": 6278, "length": 45} +{"episode_index": 6279, "length": 45} +{"episode_index": 6280, "length": 47} +{"episode_index": 6281, "length": 45} +{"episode_index": 6282, "length": 10} +{"episode_index": 6283, "length": 44} +{"episode_index": 6284, "length": 45} +{"episode_index": 6285, "length": 45} +{"episode_index": 6286, "length": 19} +{"episode_index": 6287, "length": 44} +{"episode_index": 6288, "length": 22} +{"episode_index": 6289, "length": 19} +{"episode_index": 6290, "length": 45} +{"episode_index": 6291, "length": 18} +{"episode_index": 6292, "length": 12} +{"episode_index": 6293, "length": 47} +{"episode_index": 6294, "length": 46} +{"episode_index": 6295, "length": 46} +{"episode_index": 6296, "length": 45} +{"episode_index": 6297, "length": 44} +{"episode_index": 6298, "length": 18} +{"episode_index": 6299, "length": 45} +{"episode_index": 6300, "length": 44} +{"episode_index": 6301, "length": 48} +{"episode_index": 6302, "length": 45} +{"episode_index": 6303, "length": 45} +{"episode_index": 6304, "length": 44} +{"episode_index": 6305, "length": 44} +{"episode_index": 6306, "length": 19} +{"episode_index": 6307, "length": 45} +{"episode_index": 6308, "length": 26} +{"episode_index": 6309, "length": 46} +{"episode_index": 6310, "length": 45} +{"episode_index": 6311, "length": 45} +{"episode_index": 6312, "length": 45} +{"episode_index": 6313, "length": 48} +{"episode_index": 6314, "length": 45} +{"episode_index": 6315, "length": 47} +{"episode_index": 6316, "length": 45} +{"episode_index": 6317, "length": 42} +{"episode_index": 6318, "length": 44} +{"episode_index": 6319, "length": 46} +{"episode_index": 6320, "length": 45} +{"episode_index": 6321, "length": 43} +{"episode_index": 6322, "length": 44} +{"episode_index": 6323, "length": 25} +{"episode_index": 6324, "length": 45} +{"episode_index": 6325, "length": 44} +{"episode_index": 6326, "length": 23} +{"episode_index": 6327, "length": 44} +{"episode_index": 6328, "length": 19} +{"episode_index": 6329, "length": 45} +{"episode_index": 6330, "length": 45} +{"episode_index": 6331, "length": 47} +{"episode_index": 6332, "length": 20} +{"episode_index": 6333, "length": 45} +{"episode_index": 6334, "length": 28} +{"episode_index": 6335, "length": 45} +{"episode_index": 6336, "length": 45} +{"episode_index": 6337, "length": 45} +{"episode_index": 6338, "length": 44} +{"episode_index": 6339, "length": 45} +{"episode_index": 6340, "length": 12} +{"episode_index": 6341, "length": 45} +{"episode_index": 6342, "length": 45} +{"episode_index": 6343, "length": 35} +{"episode_index": 6344, "length": 47} +{"episode_index": 6345, "length": 45} +{"episode_index": 6346, "length": 44} +{"episode_index": 6347, "length": 44} +{"episode_index": 6348, "length": 46} +{"episode_index": 6349, "length": 12} +{"episode_index": 6350, "length": 45} +{"episode_index": 6351, "length": 30} +{"episode_index": 6352, "length": 45} +{"episode_index": 6353, "length": 45} +{"episode_index": 6354, "length": 45} +{"episode_index": 6355, "length": 45} +{"episode_index": 6356, "length": 48} +{"episode_index": 6357, "length": 46} +{"episode_index": 6358, "length": 47} +{"episode_index": 6359, "length": 43} +{"episode_index": 6360, "length": 16} +{"episode_index": 6361, "length": 46} +{"episode_index": 6362, "length": 49} +{"episode_index": 6363, "length": 44} +{"episode_index": 6364, "length": 18} +{"episode_index": 6365, "length": 19} +{"episode_index": 6366, "length": 44} +{"episode_index": 6367, "length": 45} +{"episode_index": 6368, "length": 25} +{"episode_index": 6369, "length": 44} +{"episode_index": 6370, "length": 45} +{"episode_index": 6371, "length": 45} +{"episode_index": 6372, "length": 20} +{"episode_index": 6373, "length": 44} +{"episode_index": 6374, "length": 47} +{"episode_index": 6375, "length": 45} +{"episode_index": 6376, "length": 45} +{"episode_index": 6377, "length": 44} +{"episode_index": 6378, "length": 13} +{"episode_index": 6379, "length": 45} +{"episode_index": 6380, "length": 32} +{"episode_index": 6381, "length": 45} +{"episode_index": 6382, "length": 45} +{"episode_index": 6383, "length": 45} +{"episode_index": 6384, "length": 44} +{"episode_index": 6385, "length": 44} +{"episode_index": 6386, "length": 45} +{"episode_index": 6387, "length": 44} +{"episode_index": 6388, "length": 49} +{"episode_index": 6389, "length": 44} +{"episode_index": 6390, "length": 18} +{"episode_index": 6391, "length": 45} +{"episode_index": 6392, "length": 45} +{"episode_index": 6393, "length": 46} +{"episode_index": 6394, "length": 43} +{"episode_index": 6395, "length": 45} +{"episode_index": 6396, "length": 43} +{"episode_index": 6397, "length": 45} +{"episode_index": 6398, "length": 47} +{"episode_index": 6399, "length": 47} +{"episode_index": 6400, "length": 44} +{"episode_index": 6401, "length": 46} +{"episode_index": 6402, "length": 47} +{"episode_index": 6403, "length": 49} +{"episode_index": 6404, "length": 44} +{"episode_index": 6405, "length": 46} +{"episode_index": 6406, "length": 44} +{"episode_index": 6407, "length": 44} +{"episode_index": 6408, "length": 12} +{"episode_index": 6409, "length": 45} +{"episode_index": 6410, "length": 45} +{"episode_index": 6411, "length": 45} +{"episode_index": 6412, "length": 20} +{"episode_index": 6413, "length": 46} +{"episode_index": 6414, "length": 22} +{"episode_index": 6415, "length": 46} +{"episode_index": 6416, "length": 45} +{"episode_index": 6417, "length": 46} +{"episode_index": 6418, "length": 45} +{"episode_index": 6419, "length": 27} +{"episode_index": 6420, "length": 44} +{"episode_index": 6421, "length": 44} +{"episode_index": 6422, "length": 45} +{"episode_index": 6423, "length": 45} +{"episode_index": 6424, "length": 45} +{"episode_index": 6425, "length": 47} +{"episode_index": 6426, "length": 49} +{"episode_index": 6427, "length": 43} +{"episode_index": 6428, "length": 44} +{"episode_index": 6429, "length": 45} +{"episode_index": 6430, "length": 36} +{"episode_index": 6431, "length": 45} +{"episode_index": 6432, "length": 13} +{"episode_index": 6433, "length": 45} +{"episode_index": 6434, "length": 45} +{"episode_index": 6435, "length": 43} +{"episode_index": 6436, "length": 45} +{"episode_index": 6437, "length": 45} +{"episode_index": 6438, "length": 45} +{"episode_index": 6439, "length": 45} +{"episode_index": 6440, "length": 44} +{"episode_index": 6441, "length": 46} +{"episode_index": 6442, "length": 44} +{"episode_index": 6443, "length": 47} +{"episode_index": 6444, "length": 45} +{"episode_index": 6445, "length": 44} +{"episode_index": 6446, "length": 45} +{"episode_index": 6447, "length": 47} +{"episode_index": 6448, "length": 44} +{"episode_index": 6449, "length": 45} +{"episode_index": 6450, "length": 46} +{"episode_index": 6451, "length": 44} +{"episode_index": 6452, "length": 16} +{"episode_index": 6453, "length": 45} +{"episode_index": 6454, "length": 44} +{"episode_index": 6455, "length": 43} +{"episode_index": 6456, "length": 44} +{"episode_index": 6457, "length": 46} +{"episode_index": 6458, "length": 18} +{"episode_index": 6459, "length": 45} +{"episode_index": 6460, "length": 48} +{"episode_index": 6461, "length": 45} +{"episode_index": 6462, "length": 45} +{"episode_index": 6463, "length": 44} +{"episode_index": 6464, "length": 44} +{"episode_index": 6465, "length": 44} +{"episode_index": 6466, "length": 47} +{"episode_index": 6467, "length": 45} +{"episode_index": 6468, "length": 45} +{"episode_index": 6469, "length": 45} +{"episode_index": 6470, "length": 45} +{"episode_index": 6471, "length": 46} +{"episode_index": 6472, "length": 45} +{"episode_index": 6473, "length": 44} +{"episode_index": 6474, "length": 45} +{"episode_index": 6475, "length": 44} +{"episode_index": 6476, "length": 46} +{"episode_index": 6477, "length": 43} +{"episode_index": 6478, "length": 45} +{"episode_index": 6479, "length": 46} +{"episode_index": 6480, "length": 19} +{"episode_index": 6481, "length": 44} +{"episode_index": 6482, "length": 45} +{"episode_index": 6483, "length": 45} +{"episode_index": 6484, "length": 45} +{"episode_index": 6485, "length": 45} +{"episode_index": 6486, "length": 45} +{"episode_index": 6487, "length": 44} +{"episode_index": 6488, "length": 19} +{"episode_index": 6489, "length": 45} +{"episode_index": 6490, "length": 45} +{"episode_index": 6491, "length": 45} +{"episode_index": 6492, "length": 44} +{"episode_index": 6493, "length": 18} +{"episode_index": 6494, "length": 46} +{"episode_index": 6495, "length": 44} +{"episode_index": 6496, "length": 18} +{"episode_index": 6497, "length": 45} +{"episode_index": 6498, "length": 45} +{"episode_index": 6499, "length": 46} +{"episode_index": 6500, "length": 45} +{"episode_index": 6501, "length": 11} +{"episode_index": 6502, "length": 45} +{"episode_index": 6503, "length": 43} +{"episode_index": 6504, "length": 49} +{"episode_index": 6505, "length": 18} +{"episode_index": 6506, "length": 11} +{"episode_index": 6507, "length": 45} +{"episode_index": 6508, "length": 45} +{"episode_index": 6509, "length": 46} +{"episode_index": 6510, "length": 11} +{"episode_index": 6511, "length": 45} +{"episode_index": 6512, "length": 44} +{"episode_index": 6513, "length": 45} +{"episode_index": 6514, "length": 45} +{"episode_index": 6515, "length": 47} +{"episode_index": 6516, "length": 45} +{"episode_index": 6517, "length": 44} +{"episode_index": 6518, "length": 46} +{"episode_index": 6519, "length": 46} +{"episode_index": 6520, "length": 45} +{"episode_index": 6521, "length": 47} +{"episode_index": 6522, "length": 46} +{"episode_index": 6523, "length": 44} +{"episode_index": 6524, "length": 47} +{"episode_index": 6525, "length": 46} +{"episode_index": 6526, "length": 44} +{"episode_index": 6527, "length": 12} +{"episode_index": 6528, "length": 45} +{"episode_index": 6529, "length": 45} +{"episode_index": 6530, "length": 45} +{"episode_index": 6531, "length": 45} +{"episode_index": 6532, "length": 47} +{"episode_index": 6533, "length": 43} +{"episode_index": 6534, "length": 45} +{"episode_index": 6535, "length": 19} +{"episode_index": 6536, "length": 22} +{"episode_index": 6537, "length": 44} +{"episode_index": 6538, "length": 45} +{"episode_index": 6539, "length": 47} +{"episode_index": 6540, "length": 44} +{"episode_index": 6541, "length": 45} +{"episode_index": 6542, "length": 23} +{"episode_index": 6543, "length": 44} +{"episode_index": 6544, "length": 45} +{"episode_index": 6545, "length": 47} +{"episode_index": 6546, "length": 44} +{"episode_index": 6547, "length": 45} +{"episode_index": 6548, "length": 43} +{"episode_index": 6549, "length": 45} +{"episode_index": 6550, "length": 45} +{"episode_index": 6551, "length": 13} +{"episode_index": 6552, "length": 45} +{"episode_index": 6553, "length": 44} +{"episode_index": 6554, "length": 45} +{"episode_index": 6555, "length": 44} +{"episode_index": 6556, "length": 45} +{"episode_index": 6557, "length": 45} +{"episode_index": 6558, "length": 45} +{"episode_index": 6559, "length": 45} +{"episode_index": 6560, "length": 18} +{"episode_index": 6561, "length": 46} +{"episode_index": 6562, "length": 20} +{"episode_index": 6563, "length": 45} +{"episode_index": 6564, "length": 49} +{"episode_index": 6565, "length": 48} +{"episode_index": 6566, "length": 46} +{"episode_index": 6567, "length": 43} +{"episode_index": 6568, "length": 20} +{"episode_index": 6569, "length": 48} +{"episode_index": 6570, "length": 44} +{"episode_index": 6571, "length": 45} +{"episode_index": 6572, "length": 45} +{"episode_index": 6573, "length": 46} +{"episode_index": 6574, "length": 45} +{"episode_index": 6575, "length": 20} +{"episode_index": 6576, "length": 45} +{"episode_index": 6577, "length": 46} +{"episode_index": 6578, "length": 46} +{"episode_index": 6579, "length": 45} +{"episode_index": 6580, "length": 46} +{"episode_index": 6581, "length": 44} +{"episode_index": 6582, "length": 45} +{"episode_index": 6583, "length": 43} +{"episode_index": 6584, "length": 10} +{"episode_index": 6585, "length": 45} +{"episode_index": 6586, "length": 45} +{"episode_index": 6587, "length": 30} +{"episode_index": 6588, "length": 45} +{"episode_index": 6589, "length": 45} +{"episode_index": 6590, "length": 45} +{"episode_index": 6591, "length": 45} +{"episode_index": 6592, "length": 49} +{"episode_index": 6593, "length": 47} +{"episode_index": 6594, "length": 12} +{"episode_index": 6595, "length": 35} +{"episode_index": 6596, "length": 47} +{"episode_index": 6597, "length": 45} +{"episode_index": 6598, "length": 45} +{"episode_index": 6599, "length": 45} +{"episode_index": 6600, "length": 45} +{"episode_index": 6601, "length": 15} +{"episode_index": 6602, "length": 46} +{"episode_index": 6603, "length": 45} +{"episode_index": 6604, "length": 24} +{"episode_index": 6605, "length": 19} +{"episode_index": 6606, "length": 43} +{"episode_index": 6607, "length": 47} +{"episode_index": 6608, "length": 18} +{"episode_index": 6609, "length": 44} +{"episode_index": 6610, "length": 44} +{"episode_index": 6611, "length": 19} +{"episode_index": 6612, "length": 44} +{"episode_index": 6613, "length": 18} +{"episode_index": 6614, "length": 45} +{"episode_index": 6615, "length": 45} +{"episode_index": 6616, "length": 44} +{"episode_index": 6617, "length": 19} +{"episode_index": 6618, "length": 46} +{"episode_index": 6619, "length": 44} +{"episode_index": 6620, "length": 48} +{"episode_index": 6621, "length": 45} +{"episode_index": 6622, "length": 12} +{"episode_index": 6623, "length": 17} +{"episode_index": 6624, "length": 45} +{"episode_index": 6625, "length": 44} +{"episode_index": 6626, "length": 45} +{"episode_index": 6627, "length": 45} +{"episode_index": 6628, "length": 45} +{"episode_index": 6629, "length": 23} +{"episode_index": 6630, "length": 45} +{"episode_index": 6631, "length": 45} +{"episode_index": 6632, "length": 44} +{"episode_index": 6633, "length": 44} +{"episode_index": 6634, "length": 32} +{"episode_index": 6635, "length": 19} +{"episode_index": 6636, "length": 30} +{"episode_index": 6637, "length": 46} +{"episode_index": 6638, "length": 45} +{"episode_index": 6639, "length": 46} +{"episode_index": 6640, "length": 45} +{"episode_index": 6641, "length": 44} +{"episode_index": 6642, "length": 12} +{"episode_index": 6643, "length": 45} +{"episode_index": 6644, "length": 46} +{"episode_index": 6645, "length": 44} +{"episode_index": 6646, "length": 45} +{"episode_index": 6647, "length": 47} +{"episode_index": 6648, "length": 45} +{"episode_index": 6649, "length": 46} +{"episode_index": 6650, "length": 45} +{"episode_index": 6651, "length": 44} +{"episode_index": 6652, "length": 45} +{"episode_index": 6653, "length": 46} +{"episode_index": 6654, "length": 43} +{"episode_index": 6655, "length": 43} +{"episode_index": 6656, "length": 45} +{"episode_index": 6657, "length": 44} +{"episode_index": 6658, "length": 45} +{"episode_index": 6659, "length": 49} +{"episode_index": 6660, "length": 20} +{"episode_index": 6661, "length": 17} +{"episode_index": 6662, "length": 45} +{"episode_index": 6663, "length": 44} +{"episode_index": 6664, "length": 48} +{"episode_index": 6665, "length": 47} +{"episode_index": 6666, "length": 45} +{"episode_index": 6667, "length": 45} +{"episode_index": 6668, "length": 46} +{"episode_index": 6669, "length": 44} +{"episode_index": 6670, "length": 45} +{"episode_index": 6671, "length": 44} +{"episode_index": 6672, "length": 45} +{"episode_index": 6673, "length": 47} +{"episode_index": 6674, "length": 45} +{"episode_index": 6675, "length": 45} +{"episode_index": 6676, "length": 24} +{"episode_index": 6677, "length": 46} +{"episode_index": 6678, "length": 43} +{"episode_index": 6679, "length": 44} +{"episode_index": 6680, "length": 45} +{"episode_index": 6681, "length": 35} +{"episode_index": 6682, "length": 45} +{"episode_index": 6683, "length": 45} +{"episode_index": 6684, "length": 46} +{"episode_index": 6685, "length": 45} +{"episode_index": 6686, "length": 47} +{"episode_index": 6687, "length": 18} +{"episode_index": 6688, "length": 45} +{"episode_index": 6689, "length": 45} +{"episode_index": 6690, "length": 47} +{"episode_index": 6691, "length": 45} +{"episode_index": 6692, "length": 45} +{"episode_index": 6693, "length": 12} +{"episode_index": 6694, "length": 45} +{"episode_index": 6695, "length": 20} +{"episode_index": 6696, "length": 45} +{"episode_index": 6697, "length": 43} +{"episode_index": 6698, "length": 44} +{"episode_index": 6699, "length": 43} +{"episode_index": 6700, "length": 46} +{"episode_index": 6701, "length": 45} +{"episode_index": 6702, "length": 44} +{"episode_index": 6703, "length": 18} +{"episode_index": 6704, "length": 43} +{"episode_index": 6705, "length": 11} +{"episode_index": 6706, "length": 45} +{"episode_index": 6707, "length": 19} +{"episode_index": 6708, "length": 45} +{"episode_index": 6709, "length": 44} +{"episode_index": 6710, "length": 12} +{"episode_index": 6711, "length": 44} +{"episode_index": 6712, "length": 11} +{"episode_index": 6713, "length": 45} +{"episode_index": 6714, "length": 43} +{"episode_index": 6715, "length": 43} +{"episode_index": 6716, "length": 49} +{"episode_index": 6717, "length": 44} +{"episode_index": 6718, "length": 45} +{"episode_index": 6719, "length": 45} +{"episode_index": 6720, "length": 44} +{"episode_index": 6721, "length": 44} +{"episode_index": 6722, "length": 45} +{"episode_index": 6723, "length": 45} +{"episode_index": 6724, "length": 44} +{"episode_index": 6725, "length": 48} +{"episode_index": 6726, "length": 48} +{"episode_index": 6727, "length": 45} +{"episode_index": 6728, "length": 46} +{"episode_index": 6729, "length": 45} +{"episode_index": 6730, "length": 18} +{"episode_index": 6731, "length": 45} +{"episode_index": 6732, "length": 45} +{"episode_index": 6733, "length": 45} +{"episode_index": 6734, "length": 50} +{"episode_index": 6735, "length": 46} +{"episode_index": 6736, "length": 47} +{"episode_index": 6737, "length": 46} +{"episode_index": 6738, "length": 45} +{"episode_index": 6739, "length": 45} +{"episode_index": 6740, "length": 49} +{"episode_index": 6741, "length": 45} +{"episode_index": 6742, "length": 45} +{"episode_index": 6743, "length": 44} +{"episode_index": 6744, "length": 45} +{"episode_index": 6745, "length": 46} +{"episode_index": 6746, "length": 44} +{"episode_index": 6747, "length": 45} +{"episode_index": 6748, "length": 10} +{"episode_index": 6749, "length": 45} +{"episode_index": 6750, "length": 45} +{"episode_index": 6751, "length": 45} +{"episode_index": 6752, "length": 45} +{"episode_index": 6753, "length": 46} +{"episode_index": 6754, "length": 45} +{"episode_index": 6755, "length": 47} +{"episode_index": 6756, "length": 20} +{"episode_index": 6757, "length": 46} +{"episode_index": 6758, "length": 44} +{"episode_index": 6759, "length": 45} +{"episode_index": 6760, "length": 47} +{"episode_index": 6761, "length": 45} +{"episode_index": 6762, "length": 45} +{"episode_index": 6763, "length": 44} +{"episode_index": 6764, "length": 46} +{"episode_index": 6765, "length": 45} +{"episode_index": 6766, "length": 45} +{"episode_index": 6767, "length": 44} +{"episode_index": 6768, "length": 12} +{"episode_index": 6769, "length": 45} +{"episode_index": 6770, "length": 43} +{"episode_index": 6771, "length": 47} +{"episode_index": 6772, "length": 45} +{"episode_index": 6773, "length": 45} +{"episode_index": 6774, "length": 45} +{"episode_index": 6775, "length": 45} +{"episode_index": 6776, "length": 44} +{"episode_index": 6777, "length": 45} +{"episode_index": 6778, "length": 45} +{"episode_index": 6779, "length": 45} +{"episode_index": 6780, "length": 21} +{"episode_index": 6781, "length": 34} +{"episode_index": 6782, "length": 46} +{"episode_index": 6783, "length": 45} +{"episode_index": 6784, "length": 45} +{"episode_index": 6785, "length": 45} +{"episode_index": 6786, "length": 45} +{"episode_index": 6787, "length": 49} +{"episode_index": 6788, "length": 45} +{"episode_index": 6789, "length": 5} +{"episode_index": 6790, "length": 26} +{"episode_index": 6791, "length": 48} +{"episode_index": 6792, "length": 45} +{"episode_index": 6793, "length": 18} +{"episode_index": 6794, "length": 45} +{"episode_index": 6795, "length": 45} +{"episode_index": 6796, "length": 35} +{"episode_index": 6797, "length": 46} +{"episode_index": 6798, "length": 43} +{"episode_index": 6799, "length": 45} +{"episode_index": 6800, "length": 44} +{"episode_index": 6801, "length": 45} +{"episode_index": 6802, "length": 45} +{"episode_index": 6803, "length": 44} +{"episode_index": 6804, "length": 44} +{"episode_index": 6805, "length": 45} +{"episode_index": 6806, "length": 47} +{"episode_index": 6807, "length": 19} +{"episode_index": 6808, "length": 45} +{"episode_index": 6809, "length": 45} +{"episode_index": 6810, "length": 45} +{"episode_index": 6811, "length": 44} +{"episode_index": 6812, "length": 44} +{"episode_index": 6813, "length": 43} +{"episode_index": 6814, "length": 43} +{"episode_index": 6815, "length": 46} +{"episode_index": 6816, "length": 45} +{"episode_index": 6817, "length": 47} +{"episode_index": 6818, "length": 45} +{"episode_index": 6819, "length": 45} +{"episode_index": 6820, "length": 42} +{"episode_index": 6821, "length": 45} +{"episode_index": 6822, "length": 19} +{"episode_index": 6823, "length": 45} +{"episode_index": 6824, "length": 18} +{"episode_index": 6825, "length": 45} +{"episode_index": 6826, "length": 45} +{"episode_index": 6827, "length": 44} +{"episode_index": 6828, "length": 43} +{"episode_index": 6829, "length": 46} +{"episode_index": 6830, "length": 19} +{"episode_index": 6831, "length": 4} +{"episode_index": 6832, "length": 43} +{"episode_index": 6833, "length": 45} +{"episode_index": 6834, "length": 45} +{"episode_index": 6835, "length": 46} +{"episode_index": 6836, "length": 45} +{"episode_index": 6837, "length": 45} +{"episode_index": 6838, "length": 48} +{"episode_index": 6839, "length": 45} +{"episode_index": 6840, "length": 21} +{"episode_index": 6841, "length": 25} +{"episode_index": 6842, "length": 46} +{"episode_index": 6843, "length": 12} +{"episode_index": 6844, "length": 45} +{"episode_index": 6845, "length": 46} +{"episode_index": 6846, "length": 43} +{"episode_index": 6847, "length": 45} +{"episode_index": 6848, "length": 43} +{"episode_index": 6849, "length": 45} +{"episode_index": 6850, "length": 44} +{"episode_index": 6851, "length": 45} +{"episode_index": 6852, "length": 11} +{"episode_index": 6853, "length": 25} +{"episode_index": 6854, "length": 45} +{"episode_index": 6855, "length": 45} +{"episode_index": 6856, "length": 46} +{"episode_index": 6857, "length": 10} +{"episode_index": 6858, "length": 45} +{"episode_index": 6859, "length": 46} +{"episode_index": 6860, "length": 45} +{"episode_index": 6861, "length": 45} +{"episode_index": 6862, "length": 46} +{"episode_index": 6863, "length": 45} +{"episode_index": 6864, "length": 20} +{"episode_index": 6865, "length": 10} +{"episode_index": 6866, "length": 46} +{"episode_index": 6867, "length": 45} +{"episode_index": 6868, "length": 44} +{"episode_index": 6869, "length": 44} +{"episode_index": 6870, "length": 44} +{"episode_index": 6871, "length": 45} +{"episode_index": 6872, "length": 45} +{"episode_index": 6873, "length": 19} +{"episode_index": 6874, "length": 45} +{"episode_index": 6875, "length": 44} +{"episode_index": 6876, "length": 45} +{"episode_index": 6877, "length": 45} +{"episode_index": 6878, "length": 43} +{"episode_index": 6879, "length": 44} +{"episode_index": 6880, "length": 43} +{"episode_index": 6881, "length": 45} +{"episode_index": 6882, "length": 45} +{"episode_index": 6883, "length": 26} +{"episode_index": 6884, "length": 45} +{"episode_index": 6885, "length": 12} +{"episode_index": 6886, "length": 45} +{"episode_index": 6887, "length": 45} +{"episode_index": 6888, "length": 46} +{"episode_index": 6889, "length": 43} +{"episode_index": 6890, "length": 45} +{"episode_index": 6891, "length": 44} +{"episode_index": 6892, "length": 48} +{"episode_index": 6893, "length": 48} +{"episode_index": 6894, "length": 44} +{"episode_index": 6895, "length": 45} +{"episode_index": 6896, "length": 45} +{"episode_index": 6897, "length": 45} +{"episode_index": 6898, "length": 44} +{"episode_index": 6899, "length": 45} +{"episode_index": 6900, "length": 43} +{"episode_index": 6901, "length": 46} +{"episode_index": 6902, "length": 45} +{"episode_index": 6903, "length": 20} +{"episode_index": 6904, "length": 19} +{"episode_index": 6905, "length": 45} +{"episode_index": 6906, "length": 45} +{"episode_index": 6907, "length": 45} +{"episode_index": 6908, "length": 45} +{"episode_index": 6909, "length": 46} +{"episode_index": 6910, "length": 19} +{"episode_index": 6911, "length": 44} +{"episode_index": 6912, "length": 43} +{"episode_index": 6913, "length": 46} +{"episode_index": 6914, "length": 46} +{"episode_index": 6915, "length": 45} +{"episode_index": 6916, "length": 44} +{"episode_index": 6917, "length": 45} +{"episode_index": 6918, "length": 11} +{"episode_index": 6919, "length": 44} +{"episode_index": 6920, "length": 44} +{"episode_index": 6921, "length": 44} +{"episode_index": 6922, "length": 44} +{"episode_index": 6923, "length": 5} +{"episode_index": 6924, "length": 48} +{"episode_index": 6925, "length": 46} +{"episode_index": 6926, "length": 19} +{"episode_index": 6927, "length": 45} +{"episode_index": 6928, "length": 12} +{"episode_index": 6929, "length": 11} +{"episode_index": 6930, "length": 44} +{"episode_index": 6931, "length": 45} +{"episode_index": 6932, "length": 19} +{"episode_index": 6933, "length": 45} +{"episode_index": 6934, "length": 5} +{"episode_index": 6935, "length": 45} +{"episode_index": 6936, "length": 45} +{"episode_index": 6937, "length": 45} +{"episode_index": 6938, "length": 45} +{"episode_index": 6939, "length": 45} +{"episode_index": 6940, "length": 45} +{"episode_index": 6941, "length": 44} +{"episode_index": 6942, "length": 20} +{"episode_index": 6943, "length": 45} +{"episode_index": 6944, "length": 12} +{"episode_index": 6945, "length": 45} +{"episode_index": 6946, "length": 46} +{"episode_index": 6947, "length": 45} +{"episode_index": 6948, "length": 44} +{"episode_index": 6949, "length": 47} +{"episode_index": 6950, "length": 44} +{"episode_index": 6951, "length": 45} +{"episode_index": 6952, "length": 46} +{"episode_index": 6953, "length": 46} +{"episode_index": 6954, "length": 44} +{"episode_index": 6955, "length": 43} +{"episode_index": 6956, "length": 44} +{"episode_index": 6957, "length": 49} +{"episode_index": 6958, "length": 18} +{"episode_index": 6959, "length": 19} +{"episode_index": 6960, "length": 44} +{"episode_index": 6961, "length": 43} +{"episode_index": 6962, "length": 45} +{"episode_index": 6963, "length": 44} +{"episode_index": 6964, "length": 45} +{"episode_index": 6965, "length": 10} +{"episode_index": 6966, "length": 43} +{"episode_index": 6967, "length": 18} +{"episode_index": 6968, "length": 49} +{"episode_index": 6969, "length": 45} +{"episode_index": 6970, "length": 44} +{"episode_index": 6971, "length": 4} +{"episode_index": 6972, "length": 45} +{"episode_index": 6973, "length": 46} +{"episode_index": 6974, "length": 43} +{"episode_index": 6975, "length": 45} +{"episode_index": 6976, "length": 43} +{"episode_index": 6977, "length": 19} +{"episode_index": 6978, "length": 46} +{"episode_index": 6979, "length": 46} +{"episode_index": 6980, "length": 20} +{"episode_index": 6981, "length": 19} +{"episode_index": 6982, "length": 46} +{"episode_index": 6983, "length": 43} +{"episode_index": 6984, "length": 47} +{"episode_index": 6985, "length": 46} +{"episode_index": 6986, "length": 43} +{"episode_index": 6987, "length": 4} +{"episode_index": 6988, "length": 16} +{"episode_index": 6989, "length": 45} +{"episode_index": 6990, "length": 45} +{"episode_index": 6991, "length": 45} +{"episode_index": 6992, "length": 45} +{"episode_index": 6993, "length": 45} +{"episode_index": 6994, "length": 45} +{"episode_index": 6995, "length": 45} +{"episode_index": 6996, "length": 48} +{"episode_index": 6997, "length": 46} +{"episode_index": 6998, "length": 44} +{"episode_index": 6999, "length": 45} +{"episode_index": 7000, "length": 45} +{"episode_index": 7001, "length": 44} +{"episode_index": 7002, "length": 11} +{"episode_index": 7003, "length": 45} +{"episode_index": 7004, "length": 45} +{"episode_index": 7005, "length": 47} +{"episode_index": 7006, "length": 45} +{"episode_index": 7007, "length": 18} +{"episode_index": 7008, "length": 45} +{"episode_index": 7009, "length": 45} +{"episode_index": 7010, "length": 43} +{"episode_index": 7011, "length": 46} +{"episode_index": 7012, "length": 48} +{"episode_index": 7013, "length": 11} +{"episode_index": 7014, "length": 45} +{"episode_index": 7015, "length": 45} +{"episode_index": 7016, "length": 45} +{"episode_index": 7017, "length": 44} +{"episode_index": 7018, "length": 11} +{"episode_index": 7019, "length": 45} +{"episode_index": 7020, "length": 45} +{"episode_index": 7021, "length": 43} +{"episode_index": 7022, "length": 43} +{"episode_index": 7023, "length": 47} +{"episode_index": 7024, "length": 47} +{"episode_index": 7025, "length": 23} +{"episode_index": 7026, "length": 44} +{"episode_index": 7027, "length": 25} +{"episode_index": 7028, "length": 45} +{"episode_index": 7029, "length": 44} +{"episode_index": 7030, "length": 46} +{"episode_index": 7031, "length": 47} +{"episode_index": 7032, "length": 46} +{"episode_index": 7033, "length": 45} +{"episode_index": 7034, "length": 12} +{"episode_index": 7035, "length": 45} +{"episode_index": 7036, "length": 44} +{"episode_index": 7037, "length": 44} +{"episode_index": 7038, "length": 45} +{"episode_index": 7039, "length": 22} +{"episode_index": 7040, "length": 45} +{"episode_index": 7041, "length": 44} +{"episode_index": 7042, "length": 45} +{"episode_index": 7043, "length": 45} +{"episode_index": 7044, "length": 46} +{"episode_index": 7045, "length": 13} +{"episode_index": 7046, "length": 19} +{"episode_index": 7047, "length": 45} +{"episode_index": 7048, "length": 45} +{"episode_index": 7049, "length": 20} +{"episode_index": 7050, "length": 46} +{"episode_index": 7051, "length": 11} +{"episode_index": 7052, "length": 45} +{"episode_index": 7053, "length": 44} +{"episode_index": 7054, "length": 46} +{"episode_index": 7055, "length": 45} +{"episode_index": 7056, "length": 43} +{"episode_index": 7057, "length": 22} +{"episode_index": 7058, "length": 44} +{"episode_index": 7059, "length": 45} +{"episode_index": 7060, "length": 45} +{"episode_index": 7061, "length": 45} +{"episode_index": 7062, "length": 44} +{"episode_index": 7063, "length": 47} +{"episode_index": 7064, "length": 45} +{"episode_index": 7065, "length": 5} +{"episode_index": 7066, "length": 48} +{"episode_index": 7067, "length": 44} +{"episode_index": 7068, "length": 47} +{"episode_index": 7069, "length": 26} +{"episode_index": 7070, "length": 45} +{"episode_index": 7071, "length": 45} +{"episode_index": 7072, "length": 32} +{"episode_index": 7073, "length": 45} +{"episode_index": 7074, "length": 20} +{"episode_index": 7075, "length": 44} +{"episode_index": 7076, "length": 47} +{"episode_index": 7077, "length": 44} +{"episode_index": 7078, "length": 20} +{"episode_index": 7079, "length": 45} +{"episode_index": 7080, "length": 43} +{"episode_index": 7081, "length": 44} +{"episode_index": 7082, "length": 43} +{"episode_index": 7083, "length": 45} +{"episode_index": 7084, "length": 45} +{"episode_index": 7085, "length": 44} +{"episode_index": 7086, "length": 49} +{"episode_index": 7087, "length": 44} +{"episode_index": 7088, "length": 45} +{"episode_index": 7089, "length": 45} +{"episode_index": 7090, "length": 43} +{"episode_index": 7091, "length": 46} +{"episode_index": 7092, "length": 46} +{"episode_index": 7093, "length": 45} +{"episode_index": 7094, "length": 44} +{"episode_index": 7095, "length": 44} +{"episode_index": 7096, "length": 46} +{"episode_index": 7097, "length": 46} +{"episode_index": 7098, "length": 44} +{"episode_index": 7099, "length": 20} +{"episode_index": 7100, "length": 45} +{"episode_index": 7101, "length": 44} +{"episode_index": 7102, "length": 45} +{"episode_index": 7103, "length": 46} +{"episode_index": 7104, "length": 48} +{"episode_index": 7105, "length": 45} +{"episode_index": 7106, "length": 45} +{"episode_index": 7107, "length": 45} +{"episode_index": 7108, "length": 4} +{"episode_index": 7109, "length": 10} +{"episode_index": 7110, "length": 46} +{"episode_index": 7111, "length": 45} +{"episode_index": 7112, "length": 46} +{"episode_index": 7113, "length": 45} +{"episode_index": 7114, "length": 45} +{"episode_index": 7115, "length": 45} +{"episode_index": 7116, "length": 45} +{"episode_index": 7117, "length": 46} +{"episode_index": 7118, "length": 44} +{"episode_index": 7119, "length": 45} +{"episode_index": 7120, "length": 46} +{"episode_index": 7121, "length": 46} +{"episode_index": 7122, "length": 45} +{"episode_index": 7123, "length": 44} +{"episode_index": 7124, "length": 46} +{"episode_index": 7125, "length": 46} +{"episode_index": 7126, "length": 45} +{"episode_index": 7127, "length": 45} +{"episode_index": 7128, "length": 43} +{"episode_index": 7129, "length": 45} +{"episode_index": 7130, "length": 45} +{"episode_index": 7131, "length": 45} +{"episode_index": 7132, "length": 44} +{"episode_index": 7133, "length": 45} +{"episode_index": 7134, "length": 45} +{"episode_index": 7135, "length": 46} +{"episode_index": 7136, "length": 45} +{"episode_index": 7137, "length": 47} +{"episode_index": 7138, "length": 45} +{"episode_index": 7139, "length": 45} +{"episode_index": 7140, "length": 18} +{"episode_index": 7141, "length": 45} +{"episode_index": 7142, "length": 44} +{"episode_index": 7143, "length": 43} +{"episode_index": 7144, "length": 47} +{"episode_index": 7145, "length": 45} +{"episode_index": 7146, "length": 45} +{"episode_index": 7147, "length": 22} +{"episode_index": 7148, "length": 48} +{"episode_index": 7149, "length": 45} +{"episode_index": 7150, "length": 45} +{"episode_index": 7151, "length": 46} +{"episode_index": 7152, "length": 44} +{"episode_index": 7153, "length": 43} +{"episode_index": 7154, "length": 46} +{"episode_index": 7155, "length": 45} +{"episode_index": 7156, "length": 45} +{"episode_index": 7157, "length": 45} +{"episode_index": 7158, "length": 45} +{"episode_index": 7159, "length": 19} +{"episode_index": 7160, "length": 19} +{"episode_index": 7161, "length": 45} +{"episode_index": 7162, "length": 19} +{"episode_index": 7163, "length": 47} +{"episode_index": 7164, "length": 45} +{"episode_index": 7165, "length": 43} +{"episode_index": 7166, "length": 45} +{"episode_index": 7167, "length": 45} +{"episode_index": 7168, "length": 47} +{"episode_index": 7169, "length": 45} +{"episode_index": 7170, "length": 44} +{"episode_index": 7171, "length": 5} +{"episode_index": 7172, "length": 44} +{"episode_index": 7173, "length": 45} +{"episode_index": 7174, "length": 26} +{"episode_index": 7175, "length": 45} +{"episode_index": 7176, "length": 46} +{"episode_index": 7177, "length": 24} +{"episode_index": 7178, "length": 16} +{"episode_index": 7179, "length": 45} +{"episode_index": 7180, "length": 46} +{"episode_index": 7181, "length": 45} +{"episode_index": 7182, "length": 45} +{"episode_index": 7183, "length": 45} +{"episode_index": 7184, "length": 47} +{"episode_index": 7185, "length": 45} +{"episode_index": 7186, "length": 45} +{"episode_index": 7187, "length": 45} +{"episode_index": 7188, "length": 46} +{"episode_index": 7189, "length": 45} +{"episode_index": 7190, "length": 43} +{"episode_index": 7191, "length": 45} +{"episode_index": 7192, "length": 49} +{"episode_index": 7193, "length": 45} +{"episode_index": 7194, "length": 11} +{"episode_index": 7195, "length": 44} +{"episode_index": 7196, "length": 47} +{"episode_index": 7197, "length": 45} +{"episode_index": 7198, "length": 13} +{"episode_index": 7199, "length": 44} +{"episode_index": 7200, "length": 10} +{"episode_index": 7201, "length": 46} +{"episode_index": 7202, "length": 44} +{"episode_index": 7203, "length": 19} +{"episode_index": 7204, "length": 44} +{"episode_index": 7205, "length": 49} +{"episode_index": 7206, "length": 44} +{"episode_index": 7207, "length": 47} +{"episode_index": 7208, "length": 46} +{"episode_index": 7209, "length": 45} +{"episode_index": 7210, "length": 44} +{"episode_index": 7211, "length": 45} +{"episode_index": 7212, "length": 46} +{"episode_index": 7213, "length": 45} +{"episode_index": 7214, "length": 45} +{"episode_index": 7215, "length": 45} +{"episode_index": 7216, "length": 45} +{"episode_index": 7217, "length": 45} +{"episode_index": 7218, "length": 45} +{"episode_index": 7219, "length": 44} +{"episode_index": 7220, "length": 44} +{"episode_index": 7221, "length": 45} +{"episode_index": 7222, "length": 45} +{"episode_index": 7223, "length": 44} +{"episode_index": 7224, "length": 45} +{"episode_index": 7225, "length": 10} +{"episode_index": 7226, "length": 45} +{"episode_index": 7227, "length": 45} +{"episode_index": 7228, "length": 44} +{"episode_index": 7229, "length": 24} +{"episode_index": 7230, "length": 46} +{"episode_index": 7231, "length": 45} +{"episode_index": 7232, "length": 49} +{"episode_index": 7233, "length": 45} +{"episode_index": 7234, "length": 45} +{"episode_index": 7235, "length": 46} +{"episode_index": 7236, "length": 47} +{"episode_index": 7237, "length": 43} +{"episode_index": 7238, "length": 45} +{"episode_index": 7239, "length": 45} +{"episode_index": 7240, "length": 12} +{"episode_index": 7241, "length": 45} +{"episode_index": 7242, "length": 45} +{"episode_index": 7243, "length": 12} +{"episode_index": 7244, "length": 45} +{"episode_index": 7245, "length": 45} +{"episode_index": 7246, "length": 45} +{"episode_index": 7247, "length": 47} +{"episode_index": 7248, "length": 44} +{"episode_index": 7249, "length": 45} +{"episode_index": 7250, "length": 44} +{"episode_index": 7251, "length": 12} +{"episode_index": 7252, "length": 44} +{"episode_index": 7253, "length": 43} +{"episode_index": 7254, "length": 46} +{"episode_index": 7255, "length": 44} +{"episode_index": 7256, "length": 44} +{"episode_index": 7257, "length": 45} +{"episode_index": 7258, "length": 45} +{"episode_index": 7259, "length": 45} +{"episode_index": 7260, "length": 19} +{"episode_index": 7261, "length": 44} +{"episode_index": 7262, "length": 45} +{"episode_index": 7263, "length": 45} +{"episode_index": 7264, "length": 31} +{"episode_index": 7265, "length": 44} +{"episode_index": 7266, "length": 46} +{"episode_index": 7267, "length": 45} +{"episode_index": 7268, "length": 45} +{"episode_index": 7269, "length": 43} +{"episode_index": 7270, "length": 45} +{"episode_index": 7271, "length": 45} +{"episode_index": 7272, "length": 45} +{"episode_index": 7273, "length": 32} +{"episode_index": 7274, "length": 45} +{"episode_index": 7275, "length": 46} +{"episode_index": 7276, "length": 44} +{"episode_index": 7277, "length": 44} +{"episode_index": 7278, "length": 45} +{"episode_index": 7279, "length": 44} +{"episode_index": 7280, "length": 44} +{"episode_index": 7281, "length": 45} +{"episode_index": 7282, "length": 45} +{"episode_index": 7283, "length": 44} +{"episode_index": 7284, "length": 44} +{"episode_index": 7285, "length": 44} +{"episode_index": 7286, "length": 24} +{"episode_index": 7287, "length": 20} +{"episode_index": 7288, "length": 18} +{"episode_index": 7289, "length": 46} +{"episode_index": 7290, "length": 45} +{"episode_index": 7291, "length": 45} +{"episode_index": 7292, "length": 45} +{"episode_index": 7293, "length": 49} +{"episode_index": 7294, "length": 44} +{"episode_index": 7295, "length": 48} +{"episode_index": 7296, "length": 45} +{"episode_index": 7297, "length": 45} +{"episode_index": 7298, "length": 44} +{"episode_index": 7299, "length": 47} +{"episode_index": 7300, "length": 48} +{"episode_index": 7301, "length": 46} +{"episode_index": 7302, "length": 45} +{"episode_index": 7303, "length": 46} +{"episode_index": 7304, "length": 44} +{"episode_index": 7305, "length": 46} +{"episode_index": 7306, "length": 34} +{"episode_index": 7307, "length": 11} +{"episode_index": 7308, "length": 45} +{"episode_index": 7309, "length": 45} +{"episode_index": 7310, "length": 47} +{"episode_index": 7311, "length": 18} +{"episode_index": 7312, "length": 45} +{"episode_index": 7313, "length": 45} +{"episode_index": 7314, "length": 44} +{"episode_index": 7315, "length": 45} +{"episode_index": 7316, "length": 11} +{"episode_index": 7317, "length": 45} +{"episode_index": 7318, "length": 10} +{"episode_index": 7319, "length": 45} +{"episode_index": 7320, "length": 45} +{"episode_index": 7321, "length": 47} +{"episode_index": 7322, "length": 43} +{"episode_index": 7323, "length": 45} +{"episode_index": 7324, "length": 44} +{"episode_index": 7325, "length": 45} +{"episode_index": 7326, "length": 47} +{"episode_index": 7327, "length": 44} +{"episode_index": 7328, "length": 44} +{"episode_index": 7329, "length": 44} +{"episode_index": 7330, "length": 45} +{"episode_index": 7331, "length": 19} +{"episode_index": 7332, "length": 45} +{"episode_index": 7333, "length": 43} +{"episode_index": 7334, "length": 45} +{"episode_index": 7335, "length": 46} +{"episode_index": 7336, "length": 44} +{"episode_index": 7337, "length": 45} +{"episode_index": 7338, "length": 47} +{"episode_index": 7339, "length": 11} +{"episode_index": 7340, "length": 45} +{"episode_index": 7341, "length": 19} +{"episode_index": 7342, "length": 45} +{"episode_index": 7343, "length": 33} +{"episode_index": 7344, "length": 20} +{"episode_index": 7345, "length": 45} +{"episode_index": 7346, "length": 27} +{"episode_index": 7347, "length": 46} +{"episode_index": 7348, "length": 48} +{"episode_index": 7349, "length": 46} +{"episode_index": 7350, "length": 45} +{"episode_index": 7351, "length": 47} +{"episode_index": 7352, "length": 43} +{"episode_index": 7353, "length": 45} +{"episode_index": 7354, "length": 45} +{"episode_index": 7355, "length": 44} +{"episode_index": 7356, "length": 45} +{"episode_index": 7357, "length": 45} +{"episode_index": 7358, "length": 45} +{"episode_index": 7359, "length": 45} +{"episode_index": 7360, "length": 45} +{"episode_index": 7361, "length": 11} +{"episode_index": 7362, "length": 45} +{"episode_index": 7363, "length": 24} +{"episode_index": 7364, "length": 43} +{"episode_index": 7365, "length": 44} +{"episode_index": 7366, "length": 46} +{"episode_index": 7367, "length": 45} +{"episode_index": 7368, "length": 43} +{"episode_index": 7369, "length": 47} +{"episode_index": 7370, "length": 44} +{"episode_index": 7371, "length": 49} +{"episode_index": 7372, "length": 44} +{"episode_index": 7373, "length": 45} +{"episode_index": 7374, "length": 45} +{"episode_index": 7375, "length": 49} +{"episode_index": 7376, "length": 46} +{"episode_index": 7377, "length": 19} +{"episode_index": 7378, "length": 46} +{"episode_index": 7379, "length": 45} +{"episode_index": 7380, "length": 44} +{"episode_index": 7381, "length": 20} +{"episode_index": 7382, "length": 45} +{"episode_index": 7383, "length": 34} +{"episode_index": 7384, "length": 45} +{"episode_index": 7385, "length": 19} +{"episode_index": 7386, "length": 44} +{"episode_index": 7387, "length": 44} +{"episode_index": 7388, "length": 45} +{"episode_index": 7389, "length": 45} +{"episode_index": 7390, "length": 45} +{"episode_index": 7391, "length": 45} +{"episode_index": 7392, "length": 45} +{"episode_index": 7393, "length": 45} +{"episode_index": 7394, "length": 46} +{"episode_index": 7395, "length": 39} +{"episode_index": 7396, "length": 45} +{"episode_index": 7397, "length": 45} +{"episode_index": 7398, "length": 47} +{"episode_index": 7399, "length": 46} +{"episode_index": 7400, "length": 44} +{"episode_index": 7401, "length": 46} +{"episode_index": 7402, "length": 45} +{"episode_index": 7403, "length": 46} +{"episode_index": 7404, "length": 46} +{"episode_index": 7405, "length": 46} +{"episode_index": 7406, "length": 45} +{"episode_index": 7407, "length": 45} +{"episode_index": 7408, "length": 45} +{"episode_index": 7409, "length": 45} +{"episode_index": 7410, "length": 45} +{"episode_index": 7411, "length": 20} +{"episode_index": 7412, "length": 43} +{"episode_index": 7413, "length": 45} +{"episode_index": 7414, "length": 45} +{"episode_index": 7415, "length": 45} +{"episode_index": 7416, "length": 47} +{"episode_index": 7417, "length": 45} +{"episode_index": 7418, "length": 45} +{"episode_index": 7419, "length": 45} +{"episode_index": 7420, "length": 15} +{"episode_index": 7421, "length": 45} +{"episode_index": 7422, "length": 45} +{"episode_index": 7423, "length": 44} +{"episode_index": 7424, "length": 32} +{"episode_index": 7425, "length": 11} +{"episode_index": 7426, "length": 46} +{"episode_index": 7427, "length": 46} +{"episode_index": 7428, "length": 44} +{"episode_index": 7429, "length": 45} +{"episode_index": 7430, "length": 45} +{"episode_index": 7431, "length": 43} +{"episode_index": 7432, "length": 45} +{"episode_index": 7433, "length": 44} +{"episode_index": 7434, "length": 23} +{"episode_index": 7435, "length": 44} +{"episode_index": 7436, "length": 46} +{"episode_index": 7437, "length": 48} +{"episode_index": 7438, "length": 44} +{"episode_index": 7439, "length": 45} +{"episode_index": 7440, "length": 45} +{"episode_index": 7441, "length": 45} +{"episode_index": 7442, "length": 45} +{"episode_index": 7443, "length": 44} +{"episode_index": 7444, "length": 47} +{"episode_index": 7445, "length": 47} +{"episode_index": 7446, "length": 49} +{"episode_index": 7447, "length": 45} +{"episode_index": 7448, "length": 46} +{"episode_index": 7449, "length": 18} +{"episode_index": 7450, "length": 46} +{"episode_index": 7451, "length": 12} +{"episode_index": 7452, "length": 12} +{"episode_index": 7453, "length": 47} +{"episode_index": 7454, "length": 45} +{"episode_index": 7455, "length": 18} +{"episode_index": 7456, "length": 43} +{"episode_index": 7457, "length": 45} +{"episode_index": 7458, "length": 45} +{"episode_index": 7459, "length": 45} +{"episode_index": 7460, "length": 45} +{"episode_index": 7461, "length": 45} +{"episode_index": 7462, "length": 45} +{"episode_index": 7463, "length": 45} +{"episode_index": 7464, "length": 25} +{"episode_index": 7465, "length": 46} +{"episode_index": 7466, "length": 45} +{"episode_index": 7467, "length": 46} +{"episode_index": 7468, "length": 45} +{"episode_index": 7469, "length": 45} +{"episode_index": 7470, "length": 33} +{"episode_index": 7471, "length": 45} +{"episode_index": 7472, "length": 44} +{"episode_index": 7473, "length": 45} +{"episode_index": 7474, "length": 5} +{"episode_index": 7475, "length": 43} +{"episode_index": 7476, "length": 43} +{"episode_index": 7477, "length": 12} +{"episode_index": 7478, "length": 45} +{"episode_index": 7479, "length": 11} +{"episode_index": 7480, "length": 45} +{"episode_index": 7481, "length": 20} +{"episode_index": 7482, "length": 44} +{"episode_index": 7483, "length": 45} +{"episode_index": 7484, "length": 44} +{"episode_index": 7485, "length": 45} +{"episode_index": 7486, "length": 46} +{"episode_index": 7487, "length": 50} +{"episode_index": 7488, "length": 45} +{"episode_index": 7489, "length": 45} +{"episode_index": 7490, "length": 46} +{"episode_index": 7491, "length": 27} +{"episode_index": 7492, "length": 46} +{"episode_index": 7493, "length": 45} +{"episode_index": 7494, "length": 11} +{"episode_index": 7495, "length": 45} +{"episode_index": 7496, "length": 44} +{"episode_index": 7497, "length": 44} +{"episode_index": 7498, "length": 45} +{"episode_index": 7499, "length": 43} +{"episode_index": 7500, "length": 46} +{"episode_index": 7501, "length": 45} +{"episode_index": 7502, "length": 44} +{"episode_index": 7503, "length": 45} +{"episode_index": 7504, "length": 46} +{"episode_index": 7505, "length": 43} +{"episode_index": 7506, "length": 45} +{"episode_index": 7507, "length": 44} +{"episode_index": 7508, "length": 45} +{"episode_index": 7509, "length": 45} +{"episode_index": 7510, "length": 44} +{"episode_index": 7511, "length": 46} +{"episode_index": 7512, "length": 45} +{"episode_index": 7513, "length": 45} +{"episode_index": 7514, "length": 44} +{"episode_index": 7515, "length": 44} +{"episode_index": 7516, "length": 46} +{"episode_index": 7517, "length": 45} +{"episode_index": 7518, "length": 45} +{"episode_index": 7519, "length": 46} +{"episode_index": 7520, "length": 45} +{"episode_index": 7521, "length": 43} +{"episode_index": 7522, "length": 46} +{"episode_index": 7523, "length": 12} +{"episode_index": 7524, "length": 45} +{"episode_index": 7525, "length": 46} +{"episode_index": 7526, "length": 44} +{"episode_index": 7527, "length": 47} +{"episode_index": 7528, "length": 45} +{"episode_index": 7529, "length": 45} +{"episode_index": 7530, "length": 38} +{"episode_index": 7531, "length": 45} +{"episode_index": 7532, "length": 45} +{"episode_index": 7533, "length": 20} +{"episode_index": 7534, "length": 45} +{"episode_index": 7535, "length": 45} +{"episode_index": 7536, "length": 45} +{"episode_index": 7537, "length": 48} +{"episode_index": 7538, "length": 45} +{"episode_index": 7539, "length": 47} +{"episode_index": 7540, "length": 44} +{"episode_index": 7541, "length": 45} +{"episode_index": 7542, "length": 47} +{"episode_index": 7543, "length": 45} +{"episode_index": 7544, "length": 45} +{"episode_index": 7545, "length": 45} +{"episode_index": 7546, "length": 26} +{"episode_index": 7547, "length": 19} +{"episode_index": 7548, "length": 45} +{"episode_index": 7549, "length": 45} +{"episode_index": 7550, "length": 47} +{"episode_index": 7551, "length": 45} +{"episode_index": 7552, "length": 19} +{"episode_index": 7553, "length": 45} +{"episode_index": 7554, "length": 45} +{"episode_index": 7555, "length": 45} +{"episode_index": 7556, "length": 46} +{"episode_index": 7557, "length": 45} +{"episode_index": 7558, "length": 45} +{"episode_index": 7559, "length": 10} +{"episode_index": 7560, "length": 46} +{"episode_index": 7561, "length": 43} +{"episode_index": 7562, "length": 46} +{"episode_index": 7563, "length": 45} +{"episode_index": 7564, "length": 44} +{"episode_index": 7565, "length": 44} +{"episode_index": 7566, "length": 46} +{"episode_index": 7567, "length": 45} +{"episode_index": 7568, "length": 45} +{"episode_index": 7569, "length": 45} +{"episode_index": 7570, "length": 44} +{"episode_index": 7571, "length": 46} +{"episode_index": 7572, "length": 48} +{"episode_index": 7573, "length": 45} +{"episode_index": 7574, "length": 18} +{"episode_index": 7575, "length": 48} +{"episode_index": 7576, "length": 11} +{"episode_index": 7577, "length": 43} +{"episode_index": 7578, "length": 45} +{"episode_index": 7579, "length": 45} +{"episode_index": 7580, "length": 47} +{"episode_index": 7581, "length": 45} +{"episode_index": 7582, "length": 44} +{"episode_index": 7583, "length": 45} +{"episode_index": 7584, "length": 45} +{"episode_index": 7585, "length": 44} +{"episode_index": 7586, "length": 45} +{"episode_index": 7587, "length": 19} +{"episode_index": 7588, "length": 43} +{"episode_index": 7589, "length": 46} +{"episode_index": 7590, "length": 45} +{"episode_index": 7591, "length": 44} +{"episode_index": 7592, "length": 44} +{"episode_index": 7593, "length": 45} +{"episode_index": 7594, "length": 45} +{"episode_index": 7595, "length": 45} +{"episode_index": 7596, "length": 46} +{"episode_index": 7597, "length": 45} +{"episode_index": 7598, "length": 44} +{"episode_index": 7599, "length": 45} +{"episode_index": 7600, "length": 44} +{"episode_index": 7601, "length": 49} +{"episode_index": 7602, "length": 45} +{"episode_index": 7603, "length": 45} +{"episode_index": 7604, "length": 44} +{"episode_index": 7605, "length": 43} +{"episode_index": 7606, "length": 44} +{"episode_index": 7607, "length": 23} +{"episode_index": 7608, "length": 24} +{"episode_index": 7609, "length": 45} +{"episode_index": 7610, "length": 45} +{"episode_index": 7611, "length": 47} +{"episode_index": 7612, "length": 46} +{"episode_index": 7613, "length": 17} +{"episode_index": 7614, "length": 45} +{"episode_index": 7615, "length": 19} +{"episode_index": 7616, "length": 45} +{"episode_index": 7617, "length": 45} +{"episode_index": 7618, "length": 45} +{"episode_index": 7619, "length": 12} +{"episode_index": 7620, "length": 45} +{"episode_index": 7621, "length": 45} +{"episode_index": 7622, "length": 46} +{"episode_index": 7623, "length": 44} +{"episode_index": 7624, "length": 15} +{"episode_index": 7625, "length": 44} +{"episode_index": 7626, "length": 45} +{"episode_index": 7627, "length": 43} +{"episode_index": 7628, "length": 45} +{"episode_index": 7629, "length": 17} +{"episode_index": 7630, "length": 45} +{"episode_index": 7631, "length": 45} +{"episode_index": 7632, "length": 45} +{"episode_index": 7633, "length": 26} +{"episode_index": 7634, "length": 45} +{"episode_index": 7635, "length": 47} +{"episode_index": 7636, "length": 45} +{"episode_index": 7637, "length": 45} +{"episode_index": 7638, "length": 45} +{"episode_index": 7639, "length": 45} +{"episode_index": 7640, "length": 45} +{"episode_index": 7641, "length": 45} +{"episode_index": 7642, "length": 47} +{"episode_index": 7643, "length": 45} +{"episode_index": 7644, "length": 45} +{"episode_index": 7645, "length": 47} +{"episode_index": 7646, "length": 45} +{"episode_index": 7647, "length": 45} +{"episode_index": 7648, "length": 43} +{"episode_index": 7649, "length": 45} diff --git a/latency-aware/deadly-corridor/mean_rtx3090_s42_8500ep_v1/demonstrations/conversion/expected/lerobot/meta/info.json b/latency-aware/deadly-corridor/mean_rtx3090_s42_8500ep_v1/demonstrations/conversion/expected/lerobot/meta/info.json new file mode 100644 index 0000000000000000000000000000000000000000..a15bcdfb52fb405c3b6a8a01c87a3fe483184da7 --- /dev/null +++ b/latency-aware/deadly-corridor/mean_rtx3090_s42_8500ep_v1/demonstrations/conversion/expected/lerobot/meta/info.json @@ -0,0 +1,203 @@ +{ + "codebase_version": "v2.0", + "fps": 8.75, + "total_episodes": 7650, + "total_frames": 313925, + "total_tasks": 1, + "total_videos": 0, + "total_chunks": 8, + "chunks_size": 1000, + "data_path": "data/chunk-{episode_chunk:03d}/episode_{episode_index:06d}.parquet", + "video_path": "videos/chunk-{episode_chunk:03d}/{video_key}/episode_{episode_index:06d}.mp4", + "features": { + "observation.image": { + "dtype": "image", + "shape": [ + 120, + 160, + 3 + ], + "names": [ + "height", + "width", + "channel" + ], + "video_info": { + "video.fps": 8.75 + } + }, + "observation.state": { + "dtype": "float32", + "shape": [ + 120 + ], + "names": [ + "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", + "decision_history_24", + "decision_history_25", + "decision_history_26", + "decision_history_27", + "decision_history_28", + "decision_history_29", + "decision_history_30", + "decision_history_31", + "decision_history_32", + "decision_history_33", + "decision_history_34", + "decision_history_35", + "decision_history_36", + "decision_history_37", + "decision_history_38", + "decision_history_39", + "decision_history_40", + "decision_history_41", + "decision_history_42", + "decision_history_43", + "decision_history_44", + "decision_history_45", + "decision_history_46", + "decision_history_47", + "decision_history_48", + "decision_history_49", + "decision_history_50", + "decision_history_51", + "decision_history_52", + "decision_history_53", + "decision_history_54", + "decision_history_55", + "decision_history_56", + "decision_history_57", + "decision_history_58", + "decision_history_59", + "decision_history_60", + "decision_history_61", + "decision_history_62", + "decision_history_63", + "decision_history_64", + "decision_history_65", + "decision_history_66", + "decision_history_67", + "decision_history_68", + "decision_history_69", + "decision_history_70", + "decision_history_71", + "decision_history_72", + "decision_history_73", + "decision_history_74", + "decision_history_75", + "decision_history_76", + "decision_history_77", + "decision_history_78", + "decision_history_79", + "decision_history_80", + "decision_history_81", + "decision_history_82", + "decision_history_83", + "decision_history_84", + "decision_history_85", + "decision_history_86", + "decision_history_87", + "decision_history_88", + "decision_history_89", + "decision_history_90", + "decision_history_91", + "decision_history_92", + "decision_history_93", + "decision_history_94", + "decision_history_95", + "decision_history_96", + "decision_history_97", + "decision_history_98", + "decision_history_99", + "decision_history_100", + "decision_history_101", + "decision_history_102", + "decision_history_103", + "decision_history_104", + "decision_history_105", + "decision_history_106", + "decision_history_107", + "decision_history_108", + "decision_history_109", + "decision_history_110", + "decision_history_111", + "decision_history_112", + "decision_history_113", + "decision_history_114", + "decision_history_115", + "decision_history_116", + "decision_history_117", + "decision_history_118", + "decision_history_119" + ] + }, + "action": { + "dtype": "float32", + "shape": [ + 7 + ], + "names": [ + "MOVE_FORWARD", + "MOVE_BACKWARD", + "MOVE_LEFT", + "MOVE_RIGHT", + "TURN_LEFT", + "TURN_RIGHT", + "ATTACK" + ] + }, + "timestamp": { + "dtype": "float64", + "shape": [ + 1 + ] + }, + "episode_index": { + "dtype": "int64", + "shape": [ + 1 + ] + }, + "frame_index": { + "dtype": "int64", + "shape": [ + 1 + ] + }, + "task_index": { + "dtype": "int64", + "shape": [ + 1 + ] + }, + "latency": { + "dtype": "int64", + "shape": [ + 1 + ] + } + } +} \ No newline at end of file diff --git a/latency-aware/deadly-corridor/mean_rtx3090_s42_8500ep_v1/demonstrations/conversion/expected/lerobot/meta/modality.json b/latency-aware/deadly-corridor/mean_rtx3090_s42_8500ep_v1/demonstrations/conversion/expected/lerobot/meta/modality.json new file mode 100644 index 0000000000000000000000000000000000000000..20eb753bfe2bb5477e072d2b1c4e4a40e727ba17 --- /dev/null +++ b/latency-aware/deadly-corridor/mean_rtx3090_s42_8500ep_v1/demonstrations/conversion/expected/lerobot/meta/modality.json @@ -0,0 +1,30 @@ +{ + "state": { + "game_state": { + "start": 0, + "end": 120, + "dtype": "float32", + "absolute": true, + "original_key": "observation.state" + } + }, + "action": { + "button": { + "start": 0, + "end": 7, + "dtype": "float32", + "absolute": true, + "original_key": "action" + } + }, + "video": { + "image": { + "original_key": "observation.image" + } + }, + "annotation": { + "human.action.task_description": { + "original_key": "task_index" + } + } +} diff --git a/latency-aware/deadly-corridor/mean_rtx3090_s42_8500ep_v1/demonstrations/conversion/expected/lerobot/meta/tasks.jsonl b/latency-aware/deadly-corridor/mean_rtx3090_s42_8500ep_v1/demonstrations/conversion/expected/lerobot/meta/tasks.jsonl new file mode 100644 index 0000000000000000000000000000000000000000..6809e132fd849009b77002374fd0bf78930a7b4e --- /dev/null +++ b/latency-aware/deadly-corridor/mean_rtx3090_s42_8500ep_v1/demonstrations/conversion/expected/lerobot/meta/tasks.jsonl @@ -0,0 +1 @@ +{"task": "You are playing Deadly Corridor in VizDoom. Your goal is to survive the corridor, fight enemies, and reach the green armor vest at the far end. The available actions are: MOVE_FORWARD, MOVE_BACKWARD, MOVE_LEFT, MOVE_RIGHT, TURN_LEFT, TURN_RIGHT, and ATTACK. Choose the best next action from the current visual observation.", "task_index": 0} diff --git a/latency-aware/deadly-corridor/mean_rtx3090_s42_8500ep_v1/demonstrations/conversion/expected/lerobot__val/manifest.json b/latency-aware/deadly-corridor/mean_rtx3090_s42_8500ep_v1/demonstrations/conversion/expected/lerobot__val/manifest.json new file mode 100644 index 0000000000000000000000000000000000000000..2db7346d32e8dfb8bb3dae8caae3e6a6aa6f85a3 --- /dev/null +++ b/latency-aware/deadly-corridor/mean_rtx3090_s42_8500ep_v1/demonstrations/conversion/expected/lerobot__val/manifest.json @@ -0,0 +1,402 @@ +{ + "dataset_name": "lerobot__val", + "env_name": "deadly_corridor", + "episodes": 850, + "frames": 34465, + "task_prompts": [ + "You are playing Deadly Corridor in VizDoom. Your goal is to survive the corridor, fight enemies, and reach the green armor vest at the far end. The available actions are: MOVE_FORWARD, MOVE_BACKWARD, MOVE_LEFT, MOVE_RIGHT, TURN_LEFT, TURN_RIGHT, and ATTACK. Choose the best next action from the current visual observation." + ], + "format": "starvla_lerobot_v2_decision_v1", + "source": "/home/ubuntu/lzj/mean-profiling/deadly_corridor/data/raw", + "robot_type": "rl_games_deadly_corridor", + "action_layout": "multibinary_7", + "source_action_layout": "deadly_corridor_joint_54", + "action_labels": [ + "MOVE_FORWARD", + "MOVE_BACKWARD", + "MOVE_LEFT", + "MOVE_RIGHT", + "TURN_LEFT", + "TURN_RIGHT", + "ATTACK" + ], + "action_dim": 7, + "active_action_dim": 7, + "state_source": "decision_v1", + "state_dim": 120, + "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", + "decision_history_24", + "decision_history_25", + "decision_history_26", + "decision_history_27", + "decision_history_28", + "decision_history_29", + "decision_history_30", + "decision_history_31", + "decision_history_32", + "decision_history_33", + "decision_history_34", + "decision_history_35", + "decision_history_36", + "decision_history_37", + "decision_history_38", + "decision_history_39", + "decision_history_40", + "decision_history_41", + "decision_history_42", + "decision_history_43", + "decision_history_44", + "decision_history_45", + "decision_history_46", + "decision_history_47", + "decision_history_48", + "decision_history_49", + "decision_history_50", + "decision_history_51", + "decision_history_52", + "decision_history_53", + "decision_history_54", + "decision_history_55", + "decision_history_56", + "decision_history_57", + "decision_history_58", + "decision_history_59", + "decision_history_60", + "decision_history_61", + "decision_history_62", + "decision_history_63", + "decision_history_64", + "decision_history_65", + "decision_history_66", + "decision_history_67", + "decision_history_68", + "decision_history_69", + "decision_history_70", + "decision_history_71", + "decision_history_72", + "decision_history_73", + "decision_history_74", + "decision_history_75", + "decision_history_76", + "decision_history_77", + "decision_history_78", + "decision_history_79", + "decision_history_80", + "decision_history_81", + "decision_history_82", + "decision_history_83", + "decision_history_84", + "decision_history_85", + "decision_history_86", + "decision_history_87", + "decision_history_88", + "decision_history_89", + "decision_history_90", + "decision_history_91", + "decision_history_92", + "decision_history_93", + "decision_history_94", + "decision_history_95", + "decision_history_96", + "decision_history_97", + "decision_history_98", + "decision_history_99", + "decision_history_100", + "decision_history_101", + "decision_history_102", + "decision_history_103", + "decision_history_104", + "decision_history_105", + "decision_history_106", + "decision_history_107", + "decision_history_108", + "decision_history_109", + "decision_history_110", + "decision_history_111", + "decision_history_112", + "decision_history_113", + "decision_history_114", + "decision_history_115", + "decision_history_116", + "decision_history_117", + "decision_history_118", + "decision_history_119" + ], + "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, + 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, + 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, + 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, + 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, + 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, + 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, + 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, + 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": 8.75, + "obs_stride_raw_frames": 4, + "latency_prompt_map_path": "/home/ubuntu/lzj/mean-profiling/deadly_corridor/data/lerobot/latency_prompt_map.json", + "custom_mixtures_path": "/home/ubuntu/lzj/mean-profiling/deadly_corridor/data/_generated_mixtures/lerobot.json", + "training_dataset_name": "lerobot" +} diff --git a/latency-aware/deadly-corridor/mean_rtx3090_s42_8500ep_v1/demonstrations/conversion/expected/lerobot__val/meta/episodes.jsonl b/latency-aware/deadly-corridor/mean_rtx3090_s42_8500ep_v1/demonstrations/conversion/expected/lerobot__val/meta/episodes.jsonl new file mode 100644 index 0000000000000000000000000000000000000000..7418d8fabb3e6912cee33668d6e8f55b2affb51d --- /dev/null +++ b/latency-aware/deadly-corridor/mean_rtx3090_s42_8500ep_v1/demonstrations/conversion/expected/lerobot__val/meta/episodes.jsonl @@ -0,0 +1,850 @@ +{"episode_index": 0, "length": 44} +{"episode_index": 1, "length": 43} +{"episode_index": 2, "length": 46} +{"episode_index": 3, "length": 44} +{"episode_index": 4, "length": 45} +{"episode_index": 5, "length": 46} +{"episode_index": 6, "length": 46} +{"episode_index": 7, "length": 43} +{"episode_index": 8, "length": 25} +{"episode_index": 9, "length": 45} +{"episode_index": 10, "length": 43} +{"episode_index": 11, "length": 43} +{"episode_index": 12, "length": 47} +{"episode_index": 13, "length": 43} +{"episode_index": 14, "length": 45} +{"episode_index": 15, "length": 44} +{"episode_index": 16, "length": 45} +{"episode_index": 17, "length": 48} +{"episode_index": 18, "length": 46} +{"episode_index": 19, "length": 43} +{"episode_index": 20, "length": 43} +{"episode_index": 21, "length": 45} +{"episode_index": 22, "length": 12} +{"episode_index": 23, "length": 47} +{"episode_index": 24, "length": 44} +{"episode_index": 25, "length": 43} +{"episode_index": 26, "length": 45} +{"episode_index": 27, "length": 45} +{"episode_index": 28, "length": 45} +{"episode_index": 29, "length": 44} +{"episode_index": 30, "length": 45} +{"episode_index": 31, "length": 25} +{"episode_index": 32, "length": 46} +{"episode_index": 33, "length": 45} +{"episode_index": 34, "length": 43} +{"episode_index": 35, "length": 45} +{"episode_index": 36, "length": 45} +{"episode_index": 37, "length": 45} +{"episode_index": 38, "length": 44} +{"episode_index": 39, "length": 45} +{"episode_index": 40, "length": 45} +{"episode_index": 41, "length": 45} +{"episode_index": 42, "length": 45} +{"episode_index": 43, "length": 45} +{"episode_index": 44, "length": 45} +{"episode_index": 45, "length": 45} +{"episode_index": 46, "length": 45} +{"episode_index": 47, "length": 47} +{"episode_index": 48, "length": 45} +{"episode_index": 49, "length": 46} +{"episode_index": 50, "length": 47} +{"episode_index": 51, "length": 47} +{"episode_index": 52, "length": 45} +{"episode_index": 53, "length": 45} +{"episode_index": 54, "length": 46} +{"episode_index": 55, "length": 45} +{"episode_index": 56, "length": 45} +{"episode_index": 57, "length": 44} +{"episode_index": 58, "length": 44} +{"episode_index": 59, "length": 44} +{"episode_index": 60, "length": 17} +{"episode_index": 61, "length": 45} +{"episode_index": 62, "length": 25} +{"episode_index": 63, "length": 20} +{"episode_index": 64, "length": 46} +{"episode_index": 65, "length": 45} +{"episode_index": 66, "length": 46} +{"episode_index": 67, "length": 45} +{"episode_index": 68, "length": 45} +{"episode_index": 69, "length": 45} +{"episode_index": 70, "length": 45} +{"episode_index": 71, "length": 44} +{"episode_index": 72, "length": 11} +{"episode_index": 73, "length": 46} +{"episode_index": 74, "length": 46} +{"episode_index": 75, "length": 45} +{"episode_index": 76, "length": 15} +{"episode_index": 77, "length": 44} +{"episode_index": 78, "length": 47} +{"episode_index": 79, "length": 45} +{"episode_index": 80, "length": 45} +{"episode_index": 81, "length": 45} +{"episode_index": 82, "length": 46} +{"episode_index": 83, "length": 47} +{"episode_index": 84, "length": 45} +{"episode_index": 85, "length": 45} +{"episode_index": 86, "length": 48} +{"episode_index": 87, "length": 48} +{"episode_index": 88, "length": 45} +{"episode_index": 89, "length": 45} +{"episode_index": 90, "length": 48} +{"episode_index": 91, "length": 46} +{"episode_index": 92, "length": 44} +{"episode_index": 93, "length": 44} +{"episode_index": 94, "length": 45} +{"episode_index": 95, "length": 18} +{"episode_index": 96, "length": 45} +{"episode_index": 97, "length": 45} +{"episode_index": 98, "length": 30} +{"episode_index": 99, "length": 44} +{"episode_index": 100, "length": 46} +{"episode_index": 101, "length": 43} +{"episode_index": 102, "length": 45} +{"episode_index": 103, "length": 43} +{"episode_index": 104, "length": 45} +{"episode_index": 105, "length": 44} +{"episode_index": 106, "length": 48} +{"episode_index": 107, "length": 46} +{"episode_index": 108, "length": 45} +{"episode_index": 109, "length": 45} +{"episode_index": 110, "length": 47} +{"episode_index": 111, "length": 19} +{"episode_index": 112, "length": 44} +{"episode_index": 113, "length": 44} +{"episode_index": 114, "length": 44} +{"episode_index": 115, "length": 45} +{"episode_index": 116, "length": 45} +{"episode_index": 117, "length": 44} +{"episode_index": 118, "length": 45} +{"episode_index": 119, "length": 44} +{"episode_index": 120, "length": 46} +{"episode_index": 121, "length": 45} +{"episode_index": 122, "length": 47} +{"episode_index": 123, "length": 46} +{"episode_index": 124, "length": 46} +{"episode_index": 125, "length": 46} +{"episode_index": 126, "length": 45} +{"episode_index": 127, "length": 18} +{"episode_index": 128, "length": 45} +{"episode_index": 129, "length": 13} +{"episode_index": 130, "length": 49} +{"episode_index": 131, "length": 45} +{"episode_index": 132, "length": 44} +{"episode_index": 133, "length": 44} +{"episode_index": 134, "length": 45} +{"episode_index": 135, "length": 44} +{"episode_index": 136, "length": 41} +{"episode_index": 137, "length": 45} +{"episode_index": 138, "length": 45} +{"episode_index": 139, "length": 47} +{"episode_index": 140, "length": 23} +{"episode_index": 141, "length": 45} +{"episode_index": 142, "length": 44} +{"episode_index": 143, "length": 23} +{"episode_index": 144, "length": 45} +{"episode_index": 145, "length": 53} +{"episode_index": 146, "length": 45} +{"episode_index": 147, "length": 18} +{"episode_index": 148, "length": 46} +{"episode_index": 149, "length": 12} +{"episode_index": 150, "length": 44} +{"episode_index": 151, "length": 45} +{"episode_index": 152, "length": 43} +{"episode_index": 153, "length": 46} +{"episode_index": 154, "length": 43} +{"episode_index": 155, "length": 12} +{"episode_index": 156, "length": 45} +{"episode_index": 157, "length": 44} +{"episode_index": 158, "length": 45} +{"episode_index": 159, "length": 44} +{"episode_index": 160, "length": 45} +{"episode_index": 161, "length": 45} +{"episode_index": 162, "length": 45} +{"episode_index": 163, "length": 45} +{"episode_index": 164, "length": 45} +{"episode_index": 165, "length": 45} +{"episode_index": 166, "length": 45} +{"episode_index": 167, "length": 20} +{"episode_index": 168, "length": 19} +{"episode_index": 169, "length": 47} +{"episode_index": 170, "length": 15} +{"episode_index": 171, "length": 45} +{"episode_index": 172, "length": 46} +{"episode_index": 173, "length": 45} +{"episode_index": 174, "length": 44} +{"episode_index": 175, "length": 43} +{"episode_index": 176, "length": 45} +{"episode_index": 177, "length": 46} +{"episode_index": 178, "length": 44} +{"episode_index": 179, "length": 43} +{"episode_index": 180, "length": 28} +{"episode_index": 181, "length": 46} +{"episode_index": 182, "length": 47} +{"episode_index": 183, "length": 43} +{"episode_index": 184, "length": 45} +{"episode_index": 185, "length": 12} +{"episode_index": 186, "length": 46} +{"episode_index": 187, "length": 11} +{"episode_index": 188, "length": 47} +{"episode_index": 189, "length": 48} +{"episode_index": 190, "length": 45} +{"episode_index": 191, "length": 45} +{"episode_index": 192, "length": 48} +{"episode_index": 193, "length": 19} +{"episode_index": 194, "length": 45} +{"episode_index": 195, "length": 45} +{"episode_index": 196, "length": 47} +{"episode_index": 197, "length": 45} +{"episode_index": 198, "length": 44} +{"episode_index": 199, "length": 44} +{"episode_index": 200, "length": 5} +{"episode_index": 201, "length": 44} +{"episode_index": 202, "length": 45} +{"episode_index": 203, "length": 46} +{"episode_index": 204, "length": 45} +{"episode_index": 205, "length": 45} +{"episode_index": 206, "length": 45} +{"episode_index": 207, "length": 43} +{"episode_index": 208, "length": 44} +{"episode_index": 209, "length": 45} +{"episode_index": 210, "length": 44} +{"episode_index": 211, "length": 44} +{"episode_index": 212, "length": 44} +{"episode_index": 213, "length": 32} +{"episode_index": 214, "length": 45} +{"episode_index": 215, "length": 19} +{"episode_index": 216, "length": 11} +{"episode_index": 217, "length": 48} +{"episode_index": 218, "length": 44} +{"episode_index": 219, "length": 45} +{"episode_index": 220, "length": 45} +{"episode_index": 221, "length": 44} +{"episode_index": 222, "length": 47} +{"episode_index": 223, "length": 45} +{"episode_index": 224, "length": 45} +{"episode_index": 225, "length": 44} +{"episode_index": 226, "length": 44} +{"episode_index": 227, "length": 45} +{"episode_index": 228, "length": 45} +{"episode_index": 229, "length": 27} +{"episode_index": 230, "length": 45} +{"episode_index": 231, "length": 46} +{"episode_index": 232, "length": 47} +{"episode_index": 233, "length": 45} +{"episode_index": 234, "length": 19} +{"episode_index": 235, "length": 46} +{"episode_index": 236, "length": 11} +{"episode_index": 237, "length": 44} +{"episode_index": 238, "length": 44} +{"episode_index": 239, "length": 45} +{"episode_index": 240, "length": 47} +{"episode_index": 241, "length": 44} +{"episode_index": 242, "length": 11} +{"episode_index": 243, "length": 20} +{"episode_index": 244, "length": 46} +{"episode_index": 245, "length": 44} +{"episode_index": 246, "length": 45} +{"episode_index": 247, "length": 44} +{"episode_index": 248, "length": 19} +{"episode_index": 249, "length": 44} +{"episode_index": 250, "length": 46} +{"episode_index": 251, "length": 45} +{"episode_index": 252, "length": 45} +{"episode_index": 253, "length": 47} +{"episode_index": 254, "length": 45} +{"episode_index": 255, "length": 45} +{"episode_index": 256, "length": 12} +{"episode_index": 257, "length": 44} +{"episode_index": 258, "length": 45} +{"episode_index": 259, "length": 12} +{"episode_index": 260, "length": 29} +{"episode_index": 261, "length": 50} +{"episode_index": 262, "length": 44} +{"episode_index": 263, "length": 45} +{"episode_index": 264, "length": 19} +{"episode_index": 265, "length": 46} +{"episode_index": 266, "length": 18} +{"episode_index": 267, "length": 45} +{"episode_index": 268, "length": 45} +{"episode_index": 269, "length": 45} +{"episode_index": 270, "length": 48} +{"episode_index": 271, "length": 44} +{"episode_index": 272, "length": 45} +{"episode_index": 273, "length": 44} +{"episode_index": 274, "length": 16} +{"episode_index": 275, "length": 45} +{"episode_index": 276, "length": 45} +{"episode_index": 277, "length": 45} +{"episode_index": 278, "length": 48} +{"episode_index": 279, "length": 47} +{"episode_index": 280, "length": 44} +{"episode_index": 281, "length": 45} +{"episode_index": 282, "length": 45} +{"episode_index": 283, "length": 46} +{"episode_index": 284, "length": 45} +{"episode_index": 285, "length": 45} +{"episode_index": 286, "length": 45} +{"episode_index": 287, "length": 45} +{"episode_index": 288, "length": 44} +{"episode_index": 289, "length": 46} +{"episode_index": 290, "length": 45} +{"episode_index": 291, "length": 44} +{"episode_index": 292, "length": 44} +{"episode_index": 293, "length": 46} +{"episode_index": 294, "length": 19} +{"episode_index": 295, "length": 24} +{"episode_index": 296, "length": 45} +{"episode_index": 297, "length": 22} +{"episode_index": 298, "length": 47} +{"episode_index": 299, "length": 44} +{"episode_index": 300, "length": 45} +{"episode_index": 301, "length": 45} +{"episode_index": 302, "length": 45} +{"episode_index": 303, "length": 46} +{"episode_index": 304, "length": 46} +{"episode_index": 305, "length": 45} +{"episode_index": 306, "length": 45} +{"episode_index": 307, "length": 46} +{"episode_index": 308, "length": 46} +{"episode_index": 309, "length": 44} +{"episode_index": 310, "length": 18} +{"episode_index": 311, "length": 44} +{"episode_index": 312, "length": 47} +{"episode_index": 313, "length": 44} +{"episode_index": 314, "length": 44} +{"episode_index": 315, "length": 45} +{"episode_index": 316, "length": 46} +{"episode_index": 317, "length": 45} +{"episode_index": 318, "length": 45} +{"episode_index": 319, "length": 45} +{"episode_index": 320, "length": 45} +{"episode_index": 321, "length": 46} +{"episode_index": 322, "length": 27} +{"episode_index": 323, "length": 44} +{"episode_index": 324, "length": 44} +{"episode_index": 325, "length": 45} +{"episode_index": 326, "length": 44} +{"episode_index": 327, "length": 45} +{"episode_index": 328, "length": 45} +{"episode_index": 329, "length": 43} +{"episode_index": 330, "length": 46} +{"episode_index": 331, "length": 43} +{"episode_index": 332, "length": 11} +{"episode_index": 333, "length": 46} +{"episode_index": 334, "length": 45} +{"episode_index": 335, "length": 45} +{"episode_index": 336, "length": 43} +{"episode_index": 337, "length": 45} +{"episode_index": 338, "length": 46} +{"episode_index": 339, "length": 45} +{"episode_index": 340, "length": 43} +{"episode_index": 341, "length": 26} +{"episode_index": 342, "length": 45} +{"episode_index": 343, "length": 44} +{"episode_index": 344, "length": 43} +{"episode_index": 345, "length": 46} +{"episode_index": 346, "length": 46} +{"episode_index": 347, "length": 45} +{"episode_index": 348, "length": 47} +{"episode_index": 349, "length": 45} +{"episode_index": 350, "length": 45} +{"episode_index": 351, "length": 43} +{"episode_index": 352, "length": 47} +{"episode_index": 353, "length": 46} +{"episode_index": 354, "length": 47} +{"episode_index": 355, "length": 44} +{"episode_index": 356, "length": 43} +{"episode_index": 357, "length": 30} +{"episode_index": 358, "length": 43} +{"episode_index": 359, "length": 45} +{"episode_index": 360, "length": 35} +{"episode_index": 361, "length": 44} +{"episode_index": 362, "length": 45} +{"episode_index": 363, "length": 45} +{"episode_index": 364, "length": 44} +{"episode_index": 365, "length": 19} +{"episode_index": 366, "length": 45} +{"episode_index": 367, "length": 26} +{"episode_index": 368, "length": 45} +{"episode_index": 369, "length": 13} +{"episode_index": 370, "length": 46} +{"episode_index": 371, "length": 44} +{"episode_index": 372, "length": 47} +{"episode_index": 373, "length": 43} +{"episode_index": 374, "length": 45} +{"episode_index": 375, "length": 44} +{"episode_index": 376, "length": 43} +{"episode_index": 377, "length": 50} +{"episode_index": 378, "length": 47} +{"episode_index": 379, "length": 34} +{"episode_index": 380, "length": 43} +{"episode_index": 381, "length": 44} +{"episode_index": 382, "length": 44} +{"episode_index": 383, "length": 47} +{"episode_index": 384, "length": 31} +{"episode_index": 385, "length": 45} +{"episode_index": 386, "length": 49} +{"episode_index": 387, "length": 45} +{"episode_index": 388, "length": 46} +{"episode_index": 389, "length": 12} +{"episode_index": 390, "length": 45} +{"episode_index": 391, "length": 43} +{"episode_index": 392, "length": 43} +{"episode_index": 393, "length": 44} +{"episode_index": 394, "length": 44} +{"episode_index": 395, "length": 45} +{"episode_index": 396, "length": 43} +{"episode_index": 397, "length": 45} +{"episode_index": 398, "length": 47} +{"episode_index": 399, "length": 45} +{"episode_index": 400, "length": 47} +{"episode_index": 401, "length": 21} +{"episode_index": 402, "length": 45} +{"episode_index": 403, "length": 45} +{"episode_index": 404, "length": 45} +{"episode_index": 405, "length": 45} +{"episode_index": 406, "length": 46} +{"episode_index": 407, "length": 26} +{"episode_index": 408, "length": 45} +{"episode_index": 409, "length": 15} +{"episode_index": 410, "length": 45} +{"episode_index": 411, "length": 45} +{"episode_index": 412, "length": 45} +{"episode_index": 413, "length": 45} +{"episode_index": 414, "length": 48} +{"episode_index": 415, "length": 48} +{"episode_index": 416, "length": 45} +{"episode_index": 417, "length": 45} +{"episode_index": 418, "length": 44} +{"episode_index": 419, "length": 43} +{"episode_index": 420, "length": 18} +{"episode_index": 421, "length": 45} +{"episode_index": 422, "length": 45} +{"episode_index": 423, "length": 44} +{"episode_index": 424, "length": 46} +{"episode_index": 425, "length": 19} +{"episode_index": 426, "length": 45} +{"episode_index": 427, "length": 19} +{"episode_index": 428, "length": 45} +{"episode_index": 429, "length": 45} +{"episode_index": 430, "length": 47} +{"episode_index": 431, "length": 47} +{"episode_index": 432, "length": 45} +{"episode_index": 433, "length": 44} +{"episode_index": 434, "length": 43} +{"episode_index": 435, "length": 45} +{"episode_index": 436, "length": 46} +{"episode_index": 437, "length": 13} +{"episode_index": 438, "length": 44} +{"episode_index": 439, "length": 47} +{"episode_index": 440, "length": 47} +{"episode_index": 441, "length": 44} +{"episode_index": 442, "length": 45} +{"episode_index": 443, "length": 44} +{"episode_index": 444, "length": 45} +{"episode_index": 445, "length": 44} +{"episode_index": 446, "length": 45} +{"episode_index": 447, "length": 44} +{"episode_index": 448, "length": 45} +{"episode_index": 449, "length": 47} +{"episode_index": 450, "length": 43} +{"episode_index": 451, "length": 45} +{"episode_index": 452, "length": 44} +{"episode_index": 453, "length": 45} +{"episode_index": 454, "length": 45} +{"episode_index": 455, "length": 45} +{"episode_index": 456, "length": 45} +{"episode_index": 457, "length": 45} +{"episode_index": 458, "length": 45} +{"episode_index": 459, "length": 43} +{"episode_index": 460, "length": 45} +{"episode_index": 461, "length": 46} +{"episode_index": 462, "length": 43} +{"episode_index": 463, "length": 46} +{"episode_index": 464, "length": 43} +{"episode_index": 465, "length": 11} +{"episode_index": 466, "length": 23} +{"episode_index": 467, "length": 46} +{"episode_index": 468, "length": 11} +{"episode_index": 469, "length": 18} +{"episode_index": 470, "length": 26} +{"episode_index": 471, "length": 44} +{"episode_index": 472, "length": 48} +{"episode_index": 473, "length": 45} +{"episode_index": 474, "length": 45} +{"episode_index": 475, "length": 44} +{"episode_index": 476, "length": 45} +{"episode_index": 477, "length": 43} +{"episode_index": 478, "length": 46} +{"episode_index": 479, "length": 45} +{"episode_index": 480, "length": 12} +{"episode_index": 481, "length": 45} +{"episode_index": 482, "length": 45} +{"episode_index": 483, "length": 47} +{"episode_index": 484, "length": 45} +{"episode_index": 485, "length": 44} +{"episode_index": 486, "length": 45} +{"episode_index": 487, "length": 45} +{"episode_index": 488, "length": 45} +{"episode_index": 489, "length": 45} +{"episode_index": 490, "length": 46} +{"episode_index": 491, "length": 45} +{"episode_index": 492, "length": 3} +{"episode_index": 493, "length": 45} +{"episode_index": 494, "length": 13} +{"episode_index": 495, "length": 18} +{"episode_index": 496, "length": 45} +{"episode_index": 497, "length": 44} +{"episode_index": 498, "length": 45} +{"episode_index": 499, "length": 45} +{"episode_index": 500, "length": 45} +{"episode_index": 501, "length": 46} +{"episode_index": 502, "length": 45} +{"episode_index": 503, "length": 45} +{"episode_index": 504, "length": 45} +{"episode_index": 505, "length": 48} +{"episode_index": 506, "length": 45} +{"episode_index": 507, "length": 44} +{"episode_index": 508, "length": 45} +{"episode_index": 509, "length": 44} +{"episode_index": 510, "length": 46} +{"episode_index": 511, "length": 43} +{"episode_index": 512, "length": 33} +{"episode_index": 513, "length": 44} +{"episode_index": 514, "length": 11} +{"episode_index": 515, "length": 45} +{"episode_index": 516, "length": 44} +{"episode_index": 517, "length": 45} +{"episode_index": 518, "length": 44} +{"episode_index": 519, "length": 12} +{"episode_index": 520, "length": 44} +{"episode_index": 521, "length": 45} +{"episode_index": 522, "length": 45} +{"episode_index": 523, "length": 44} +{"episode_index": 524, "length": 46} +{"episode_index": 525, "length": 44} +{"episode_index": 526, "length": 45} +{"episode_index": 527, "length": 45} +{"episode_index": 528, "length": 47} +{"episode_index": 529, "length": 22} +{"episode_index": 530, "length": 43} +{"episode_index": 531, "length": 44} +{"episode_index": 532, "length": 44} +{"episode_index": 533, "length": 25} +{"episode_index": 534, "length": 46} +{"episode_index": 535, "length": 45} +{"episode_index": 536, "length": 44} +{"episode_index": 537, "length": 18} +{"episode_index": 538, "length": 45} +{"episode_index": 539, "length": 45} +{"episode_index": 540, "length": 45} +{"episode_index": 541, "length": 44} +{"episode_index": 542, "length": 20} +{"episode_index": 543, "length": 45} +{"episode_index": 544, "length": 45} +{"episode_index": 545, "length": 45} +{"episode_index": 546, "length": 45} +{"episode_index": 547, "length": 26} +{"episode_index": 548, "length": 43} +{"episode_index": 549, "length": 12} +{"episode_index": 550, "length": 45} +{"episode_index": 551, "length": 45} +{"episode_index": 552, "length": 45} +{"episode_index": 553, "length": 11} +{"episode_index": 554, "length": 44} +{"episode_index": 555, "length": 19} +{"episode_index": 556, "length": 45} +{"episode_index": 557, "length": 44} +{"episode_index": 558, "length": 44} +{"episode_index": 559, "length": 46} +{"episode_index": 560, "length": 44} +{"episode_index": 561, "length": 44} +{"episode_index": 562, "length": 47} +{"episode_index": 563, "length": 44} +{"episode_index": 564, "length": 12} +{"episode_index": 565, "length": 12} +{"episode_index": 566, "length": 46} +{"episode_index": 567, "length": 28} +{"episode_index": 568, "length": 46} +{"episode_index": 569, "length": 47} +{"episode_index": 570, "length": 11} +{"episode_index": 571, "length": 45} +{"episode_index": 572, "length": 44} +{"episode_index": 573, "length": 45} +{"episode_index": 574, "length": 45} +{"episode_index": 575, "length": 44} +{"episode_index": 576, "length": 45} +{"episode_index": 577, "length": 29} +{"episode_index": 578, "length": 43} +{"episode_index": 579, "length": 46} +{"episode_index": 580, "length": 44} +{"episode_index": 581, "length": 45} +{"episode_index": 582, "length": 46} +{"episode_index": 583, "length": 26} +{"episode_index": 584, "length": 12} +{"episode_index": 585, "length": 45} +{"episode_index": 586, "length": 10} +{"episode_index": 587, "length": 44} +{"episode_index": 588, "length": 45} +{"episode_index": 589, "length": 49} +{"episode_index": 590, "length": 18} +{"episode_index": 591, "length": 44} +{"episode_index": 592, "length": 19} +{"episode_index": 593, "length": 45} +{"episode_index": 594, "length": 11} +{"episode_index": 595, "length": 45} +{"episode_index": 596, "length": 44} +{"episode_index": 597, "length": 15} +{"episode_index": 598, "length": 45} +{"episode_index": 599, "length": 13} +{"episode_index": 600, "length": 45} +{"episode_index": 601, "length": 45} +{"episode_index": 602, "length": 45} +{"episode_index": 603, "length": 45} +{"episode_index": 604, "length": 45} +{"episode_index": 605, "length": 46} +{"episode_index": 606, "length": 43} +{"episode_index": 607, "length": 45} +{"episode_index": 608, "length": 45} +{"episode_index": 609, "length": 15} +{"episode_index": 610, "length": 46} +{"episode_index": 611, "length": 30} +{"episode_index": 612, "length": 19} +{"episode_index": 613, "length": 43} +{"episode_index": 614, "length": 44} +{"episode_index": 615, "length": 43} +{"episode_index": 616, "length": 47} +{"episode_index": 617, "length": 19} +{"episode_index": 618, "length": 45} +{"episode_index": 619, "length": 45} +{"episode_index": 620, "length": 49} +{"episode_index": 621, "length": 46} +{"episode_index": 622, "length": 12} +{"episode_index": 623, "length": 46} +{"episode_index": 624, "length": 48} +{"episode_index": 625, "length": 45} +{"episode_index": 626, "length": 50} +{"episode_index": 627, "length": 45} +{"episode_index": 628, "length": 44} +{"episode_index": 629, "length": 44} +{"episode_index": 630, "length": 18} +{"episode_index": 631, "length": 45} +{"episode_index": 632, "length": 4} +{"episode_index": 633, "length": 46} +{"episode_index": 634, "length": 46} +{"episode_index": 635, "length": 5} +{"episode_index": 636, "length": 47} +{"episode_index": 637, "length": 44} +{"episode_index": 638, "length": 47} +{"episode_index": 639, "length": 46} +{"episode_index": 640, "length": 47} +{"episode_index": 641, "length": 19} +{"episode_index": 642, "length": 44} +{"episode_index": 643, "length": 44} +{"episode_index": 644, "length": 17} +{"episode_index": 645, "length": 11} +{"episode_index": 646, "length": 43} +{"episode_index": 647, "length": 12} +{"episode_index": 648, "length": 45} +{"episode_index": 649, "length": 44} +{"episode_index": 650, "length": 45} +{"episode_index": 651, "length": 46} +{"episode_index": 652, "length": 45} +{"episode_index": 653, "length": 49} +{"episode_index": 654, "length": 48} +{"episode_index": 655, "length": 44} +{"episode_index": 656, "length": 46} +{"episode_index": 657, "length": 45} +{"episode_index": 658, "length": 48} +{"episode_index": 659, "length": 5} +{"episode_index": 660, "length": 44} +{"episode_index": 661, "length": 45} +{"episode_index": 662, "length": 43} +{"episode_index": 663, "length": 44} +{"episode_index": 664, "length": 45} +{"episode_index": 665, "length": 47} +{"episode_index": 666, "length": 45} +{"episode_index": 667, "length": 44} +{"episode_index": 668, "length": 43} +{"episode_index": 669, "length": 45} +{"episode_index": 670, "length": 18} +{"episode_index": 671, "length": 11} +{"episode_index": 672, "length": 45} +{"episode_index": 673, "length": 23} +{"episode_index": 674, "length": 20} +{"episode_index": 675, "length": 46} +{"episode_index": 676, "length": 45} +{"episode_index": 677, "length": 44} +{"episode_index": 678, "length": 24} +{"episode_index": 679, "length": 44} +{"episode_index": 680, "length": 45} +{"episode_index": 681, "length": 44} +{"episode_index": 682, "length": 44} +{"episode_index": 683, "length": 46} +{"episode_index": 684, "length": 45} +{"episode_index": 685, "length": 47} +{"episode_index": 686, "length": 45} +{"episode_index": 687, "length": 44} +{"episode_index": 688, "length": 47} +{"episode_index": 689, "length": 45} +{"episode_index": 690, "length": 45} +{"episode_index": 691, "length": 34} +{"episode_index": 692, "length": 45} +{"episode_index": 693, "length": 45} +{"episode_index": 694, "length": 44} +{"episode_index": 695, "length": 19} +{"episode_index": 696, "length": 46} +{"episode_index": 697, "length": 45} +{"episode_index": 698, "length": 46} +{"episode_index": 699, "length": 12} +{"episode_index": 700, "length": 45} +{"episode_index": 701, "length": 45} +{"episode_index": 702, "length": 18} +{"episode_index": 703, "length": 45} +{"episode_index": 704, "length": 44} +{"episode_index": 705, "length": 44} +{"episode_index": 706, "length": 45} +{"episode_index": 707, "length": 45} +{"episode_index": 708, "length": 44} +{"episode_index": 709, "length": 12} +{"episode_index": 710, "length": 18} +{"episode_index": 711, "length": 45} +{"episode_index": 712, "length": 46} +{"episode_index": 713, "length": 44} +{"episode_index": 714, "length": 45} +{"episode_index": 715, "length": 45} +{"episode_index": 716, "length": 45} +{"episode_index": 717, "length": 45} +{"episode_index": 718, "length": 46} +{"episode_index": 719, "length": 44} +{"episode_index": 720, "length": 47} +{"episode_index": 721, "length": 46} +{"episode_index": 722, "length": 46} +{"episode_index": 723, "length": 47} +{"episode_index": 724, "length": 45} +{"episode_index": 725, "length": 39} +{"episode_index": 726, "length": 45} +{"episode_index": 727, "length": 45} +{"episode_index": 728, "length": 45} +{"episode_index": 729, "length": 20} +{"episode_index": 730, "length": 43} +{"episode_index": 731, "length": 46} +{"episode_index": 732, "length": 18} +{"episode_index": 733, "length": 47} +{"episode_index": 734, "length": 45} +{"episode_index": 735, "length": 43} +{"episode_index": 736, "length": 12} +{"episode_index": 737, "length": 44} +{"episode_index": 738, "length": 45} +{"episode_index": 739, "length": 46} +{"episode_index": 740, "length": 18} +{"episode_index": 741, "length": 45} +{"episode_index": 742, "length": 44} +{"episode_index": 743, "length": 44} +{"episode_index": 744, "length": 47} +{"episode_index": 745, "length": 11} +{"episode_index": 746, "length": 45} +{"episode_index": 747, "length": 45} +{"episode_index": 748, "length": 43} +{"episode_index": 749, "length": 44} +{"episode_index": 750, "length": 45} +{"episode_index": 751, "length": 44} +{"episode_index": 752, "length": 44} +{"episode_index": 753, "length": 45} +{"episode_index": 754, "length": 46} +{"episode_index": 755, "length": 45} +{"episode_index": 756, "length": 45} +{"episode_index": 757, "length": 45} +{"episode_index": 758, "length": 44} +{"episode_index": 759, "length": 5} +{"episode_index": 760, "length": 45} +{"episode_index": 761, "length": 45} +{"episode_index": 762, "length": 48} +{"episode_index": 763, "length": 44} +{"episode_index": 764, "length": 45} +{"episode_index": 765, "length": 44} +{"episode_index": 766, "length": 45} +{"episode_index": 767, "length": 45} +{"episode_index": 768, "length": 12} +{"episode_index": 769, "length": 33} +{"episode_index": 770, "length": 45} +{"episode_index": 771, "length": 45} +{"episode_index": 772, "length": 47} +{"episode_index": 773, "length": 17} +{"episode_index": 774, "length": 45} +{"episode_index": 775, "length": 45} +{"episode_index": 776, "length": 42} +{"episode_index": 777, "length": 48} +{"episode_index": 778, "length": 46} +{"episode_index": 779, "length": 46} +{"episode_index": 780, "length": 46} +{"episode_index": 781, "length": 47} +{"episode_index": 782, "length": 25} +{"episode_index": 783, "length": 32} +{"episode_index": 784, "length": 45} +{"episode_index": 785, "length": 45} +{"episode_index": 786, "length": 45} +{"episode_index": 787, "length": 44} +{"episode_index": 788, "length": 45} +{"episode_index": 789, "length": 45} +{"episode_index": 790, "length": 44} +{"episode_index": 791, "length": 18} +{"episode_index": 792, "length": 47} +{"episode_index": 793, "length": 46} +{"episode_index": 794, "length": 45} +{"episode_index": 795, "length": 46} +{"episode_index": 796, "length": 45} +{"episode_index": 797, "length": 46} +{"episode_index": 798, "length": 45} +{"episode_index": 799, "length": 44} +{"episode_index": 800, "length": 45} +{"episode_index": 801, "length": 45} +{"episode_index": 802, "length": 46} +{"episode_index": 803, "length": 46} +{"episode_index": 804, "length": 45} +{"episode_index": 805, "length": 45} +{"episode_index": 806, "length": 45} +{"episode_index": 807, "length": 4} +{"episode_index": 808, "length": 45} +{"episode_index": 809, "length": 11} +{"episode_index": 810, "length": 11} +{"episode_index": 811, "length": 18} +{"episode_index": 812, "length": 46} +{"episode_index": 813, "length": 45} +{"episode_index": 814, "length": 45} +{"episode_index": 815, "length": 20} +{"episode_index": 816, "length": 44} +{"episode_index": 817, "length": 15} +{"episode_index": 818, "length": 44} +{"episode_index": 819, "length": 47} +{"episode_index": 820, "length": 44} +{"episode_index": 821, "length": 22} +{"episode_index": 822, "length": 45} +{"episode_index": 823, "length": 45} +{"episode_index": 824, "length": 11} +{"episode_index": 825, "length": 44} +{"episode_index": 826, "length": 45} +{"episode_index": 827, "length": 45} +{"episode_index": 828, "length": 18} +{"episode_index": 829, "length": 45} +{"episode_index": 830, "length": 45} +{"episode_index": 831, "length": 45} +{"episode_index": 832, "length": 44} +{"episode_index": 833, "length": 45} +{"episode_index": 834, "length": 45} +{"episode_index": 835, "length": 47} +{"episode_index": 836, "length": 45} +{"episode_index": 837, "length": 44} +{"episode_index": 838, "length": 43} +{"episode_index": 839, "length": 45} +{"episode_index": 840, "length": 46} +{"episode_index": 841, "length": 45} +{"episode_index": 842, "length": 43} +{"episode_index": 843, "length": 46} +{"episode_index": 844, "length": 44} +{"episode_index": 845, "length": 43} +{"episode_index": 846, "length": 45} +{"episode_index": 847, "length": 47} +{"episode_index": 848, "length": 45} +{"episode_index": 849, "length": 46} diff --git a/latency-aware/deadly-corridor/mean_rtx3090_s42_8500ep_v1/demonstrations/conversion/expected/lerobot__val/meta/info.json b/latency-aware/deadly-corridor/mean_rtx3090_s42_8500ep_v1/demonstrations/conversion/expected/lerobot__val/meta/info.json new file mode 100644 index 0000000000000000000000000000000000000000..e6bf1151858c58184e8e924a4d7c1292872c0a47 --- /dev/null +++ b/latency-aware/deadly-corridor/mean_rtx3090_s42_8500ep_v1/demonstrations/conversion/expected/lerobot__val/meta/info.json @@ -0,0 +1,203 @@ +{ + "codebase_version": "v2.0", + "fps": 8.75, + "total_episodes": 850, + "total_frames": 34465, + "total_tasks": 1, + "total_videos": 0, + "total_chunks": 1, + "chunks_size": 1000, + "data_path": "data/chunk-{episode_chunk:03d}/episode_{episode_index:06d}.parquet", + "video_path": "videos/chunk-{episode_chunk:03d}/{video_key}/episode_{episode_index:06d}.mp4", + "features": { + "observation.image": { + "dtype": "image", + "shape": [ + 120, + 160, + 3 + ], + "names": [ + "height", + "width", + "channel" + ], + "video_info": { + "video.fps": 8.75 + } + }, + "observation.state": { + "dtype": "float32", + "shape": [ + 120 + ], + "names": [ + "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", + "decision_history_24", + "decision_history_25", + "decision_history_26", + "decision_history_27", + "decision_history_28", + "decision_history_29", + "decision_history_30", + "decision_history_31", + "decision_history_32", + "decision_history_33", + "decision_history_34", + "decision_history_35", + "decision_history_36", + "decision_history_37", + "decision_history_38", + "decision_history_39", + "decision_history_40", + "decision_history_41", + "decision_history_42", + "decision_history_43", + "decision_history_44", + "decision_history_45", + "decision_history_46", + "decision_history_47", + "decision_history_48", + "decision_history_49", + "decision_history_50", + "decision_history_51", + "decision_history_52", + "decision_history_53", + "decision_history_54", + "decision_history_55", + "decision_history_56", + "decision_history_57", + "decision_history_58", + "decision_history_59", + "decision_history_60", + "decision_history_61", + "decision_history_62", + "decision_history_63", + "decision_history_64", + "decision_history_65", + "decision_history_66", + "decision_history_67", + "decision_history_68", + "decision_history_69", + "decision_history_70", + "decision_history_71", + "decision_history_72", + "decision_history_73", + "decision_history_74", + "decision_history_75", + "decision_history_76", + "decision_history_77", + "decision_history_78", + "decision_history_79", + "decision_history_80", + "decision_history_81", + "decision_history_82", + "decision_history_83", + "decision_history_84", + "decision_history_85", + "decision_history_86", + "decision_history_87", + "decision_history_88", + "decision_history_89", + "decision_history_90", + "decision_history_91", + "decision_history_92", + "decision_history_93", + "decision_history_94", + "decision_history_95", + "decision_history_96", + "decision_history_97", + "decision_history_98", + "decision_history_99", + "decision_history_100", + "decision_history_101", + "decision_history_102", + "decision_history_103", + "decision_history_104", + "decision_history_105", + "decision_history_106", + "decision_history_107", + "decision_history_108", + "decision_history_109", + "decision_history_110", + "decision_history_111", + "decision_history_112", + "decision_history_113", + "decision_history_114", + "decision_history_115", + "decision_history_116", + "decision_history_117", + "decision_history_118", + "decision_history_119" + ] + }, + "action": { + "dtype": "float32", + "shape": [ + 7 + ], + "names": [ + "MOVE_FORWARD", + "MOVE_BACKWARD", + "MOVE_LEFT", + "MOVE_RIGHT", + "TURN_LEFT", + "TURN_RIGHT", + "ATTACK" + ] + }, + "timestamp": { + "dtype": "float64", + "shape": [ + 1 + ] + }, + "episode_index": { + "dtype": "int64", + "shape": [ + 1 + ] + }, + "frame_index": { + "dtype": "int64", + "shape": [ + 1 + ] + }, + "task_index": { + "dtype": "int64", + "shape": [ + 1 + ] + }, + "latency": { + "dtype": "int64", + "shape": [ + 1 + ] + } + } +} \ No newline at end of file diff --git a/latency-aware/deadly-corridor/mean_rtx3090_s42_8500ep_v1/demonstrations/conversion/expected/lerobot__val/meta/modality.json b/latency-aware/deadly-corridor/mean_rtx3090_s42_8500ep_v1/demonstrations/conversion/expected/lerobot__val/meta/modality.json new file mode 100644 index 0000000000000000000000000000000000000000..20eb753bfe2bb5477e072d2b1c4e4a40e727ba17 --- /dev/null +++ b/latency-aware/deadly-corridor/mean_rtx3090_s42_8500ep_v1/demonstrations/conversion/expected/lerobot__val/meta/modality.json @@ -0,0 +1,30 @@ +{ + "state": { + "game_state": { + "start": 0, + "end": 120, + "dtype": "float32", + "absolute": true, + "original_key": "observation.state" + } + }, + "action": { + "button": { + "start": 0, + "end": 7, + "dtype": "float32", + "absolute": true, + "original_key": "action" + } + }, + "video": { + "image": { + "original_key": "observation.image" + } + }, + "annotation": { + "human.action.task_description": { + "original_key": "task_index" + } + } +} diff --git a/latency-aware/deadly-corridor/mean_rtx3090_s42_8500ep_v1/demonstrations/conversion/expected/lerobot__val/meta/tasks.jsonl b/latency-aware/deadly-corridor/mean_rtx3090_s42_8500ep_v1/demonstrations/conversion/expected/lerobot__val/meta/tasks.jsonl new file mode 100644 index 0000000000000000000000000000000000000000..6809e132fd849009b77002374fd0bf78930a7b4e --- /dev/null +++ b/latency-aware/deadly-corridor/mean_rtx3090_s42_8500ep_v1/demonstrations/conversion/expected/lerobot__val/meta/tasks.jsonl @@ -0,0 +1 @@ +{"task": "You are playing Deadly Corridor in VizDoom. Your goal is to survive the corridor, fight enemies, and reach the green armor vest at the far end. The available actions are: MOVE_FORWARD, MOVE_BACKWARD, MOVE_LEFT, MOVE_RIGHT, TURN_LEFT, TURN_RIGHT, and ATTACK. Choose the best next action from the current visual observation.", "task_index": 0} diff --git a/latency-aware/deadly-corridor/mean_rtx3090_s42_8500ep_v1/demonstrations/raw/filter_report.json b/latency-aware/deadly-corridor/mean_rtx3090_s42_8500ep_v1/demonstrations/raw/filter_report.json new file mode 100644 index 0000000000000000000000000000000000000000..e902cebdfd8bb5b36cd3771e56beaa1d728865f4 --- /dev/null +++ b/latency-aware/deadly-corridor/mean_rtx3090_s42_8500ep_v1/demonstrations/raw/filter_report.json @@ -0,0 +1,12 @@ +{ + "accepted_episodes": 8500, + "accepted_train_episodes": 7650, + "accepted_val_episodes": 850, + "attempted_episodes": 8500, + "preset": "none", + "rejected_episodes": 0, + "train_rows_after_row_filter": 313925, + "train_rows_before_row_filter": 313925, + "val_rows_after_row_filter": 34465, + "val_rows_before_row_filter": 34465 +} diff --git a/latency-aware/deadly-corridor/mean_rtx3090_s42_8500ep_v1/demonstrations/raw/metadata.json b/latency-aware/deadly-corridor/mean_rtx3090_s42_8500ep_v1/demonstrations/raw/metadata.json new file mode 100644 index 0000000000000000000000000000000000000000..4c29c7793d543fe136d0c94fbbd59e22e27bef48 --- /dev/null +++ b/latency-aware/deadly-corridor/mean_rtx3090_s42_8500ep_v1/demonstrations/raw/metadata.json @@ -0,0 +1,395 @@ +{ + "action_history_decisions": 8, + "action_labels": [ + "NOOP", + "ATTACK", + "MOVE_LEFT", + "MOVE_LEFT + ATTACK", + "MOVE_RIGHT", + "MOVE_RIGHT + ATTACK", + "MOVE_FORWARD", + "MOVE_FORWARD + ATTACK", + "MOVE_FORWARD + MOVE_LEFT", + "MOVE_FORWARD + MOVE_LEFT + ATTACK", + "MOVE_FORWARD + MOVE_RIGHT", + "MOVE_FORWARD + MOVE_RIGHT + ATTACK", + "MOVE_BACKWARD", + "MOVE_BACKWARD + ATTACK", + "MOVE_BACKWARD + MOVE_LEFT", + "MOVE_BACKWARD + MOVE_LEFT + ATTACK", + "MOVE_BACKWARD + MOVE_RIGHT", + "MOVE_BACKWARD + MOVE_RIGHT + ATTACK", + "TURN_LEFT", + "TURN_LEFT + ATTACK", + "TURN_LEFT + MOVE_LEFT", + "TURN_LEFT + MOVE_LEFT + ATTACK", + "TURN_LEFT + MOVE_RIGHT", + "TURN_LEFT + MOVE_RIGHT + ATTACK", + "TURN_LEFT + MOVE_FORWARD", + "TURN_LEFT + MOVE_FORWARD + ATTACK", + "TURN_LEFT + MOVE_FORWARD + MOVE_LEFT", + "TURN_LEFT + MOVE_FORWARD + MOVE_LEFT + ATTACK", + "TURN_LEFT + MOVE_FORWARD + MOVE_RIGHT", + "TURN_LEFT + MOVE_FORWARD + MOVE_RIGHT + ATTACK", + "TURN_LEFT + MOVE_BACKWARD", + "TURN_LEFT + MOVE_BACKWARD + ATTACK", + "TURN_LEFT + MOVE_BACKWARD + MOVE_LEFT", + "TURN_LEFT + MOVE_BACKWARD + MOVE_LEFT + ATTACK", + "TURN_LEFT + MOVE_BACKWARD + MOVE_RIGHT", + "TURN_LEFT + MOVE_BACKWARD + MOVE_RIGHT + ATTACK", + "TURN_RIGHT", + "TURN_RIGHT + ATTACK", + "TURN_RIGHT + MOVE_LEFT", + "TURN_RIGHT + MOVE_LEFT + ATTACK", + "TURN_RIGHT + MOVE_RIGHT", + "TURN_RIGHT + MOVE_RIGHT + ATTACK", + "TURN_RIGHT + MOVE_FORWARD", + "TURN_RIGHT + MOVE_FORWARD + ATTACK", + "TURN_RIGHT + MOVE_FORWARD + MOVE_LEFT", + "TURN_RIGHT + MOVE_FORWARD + MOVE_LEFT + ATTACK", + "TURN_RIGHT + MOVE_FORWARD + MOVE_RIGHT", + "TURN_RIGHT + MOVE_FORWARD + MOVE_RIGHT + ATTACK", + "TURN_RIGHT + MOVE_BACKWARD", + "TURN_RIGHT + MOVE_BACKWARD + ATTACK", + "TURN_RIGHT + MOVE_BACKWARD + MOVE_LEFT", + "TURN_RIGHT + MOVE_BACKWARD + MOVE_LEFT + ATTACK", + "TURN_RIGHT + MOVE_BACKWARD + MOVE_RIGHT", + "TURN_RIGHT + MOVE_BACKWARD + MOVE_RIGHT + ATTACK" + ], + "action_layout": "deadly_corridor_joint_54", + "base_prompt": "You are playing Deadly Corridor in VizDoom. Your goal is to survive the corridor, fight enemies, and reach the green armor vest at the far end. The available actions are: MOVE_FORWARD, MOVE_BACKWARD, MOVE_LEFT, MOVE_RIGHT, TURN_LEFT, TURN_RIGHT, and ATTACK. Choose the best next action from the current visual observation.", + "checkpoint_config": { + "action_history_decisions": 8, + "actor_critic_share_weights": true, + "actor_worker_gpus": [], + "adam_beta1": 0.9, + "adam_beta2": 0.999, + "adam_eps": 1e-06, + "adaptive_stddev": true, + "add_latency_info": false, + "algo": "APPO", + "async_rl": true, + "batch_size": 1024, + "batched_sampling": false, + "benchmark": false, + "cli_args": { + "action_history_decisions": 8, + "add_latency_info": false, + "algo": "APPO", + "async_rl": true, + "batch_size": 1024, + "device": "gpu", + "disable_reward_scaling": true, + "env": "doom_deadly_corridor", + "env_fps": 35.0, + "env_frameskip": 1, + "episode_metrics_path": "/home/ubuntu/lzj/mean-profiling/deadly_corridor/teacher/deadly_corridor_mean_3090_s42/episode_metrics.jsonl", + "eval_deterministic": true, + "eval_env_frameskip": 1, + "eval_episodes": 100, + "eval_max_steps": 3600, + "eval_parallel_envs": 1, + "eval_raw_reward": true, + "experiment": "deadly_corridor_mean_3090_s42", + "experiment_summaries_interval": 1, + "exploration_loss": "symmetric_kl", + "exploration_loss_coeff": 0.001, + "frame_stack": 4, + "gae_lambda": 0.95, + "gamma": 0.99, + "hold_policy": "hold", + "keep_checkpoints": 5, + "latency_profile_path": null, + "latency_profile_worker_slot": null, + "latency_seed": 42, + "latency_type": "fixed", + "learning_rate": 0.0001, + "max_grad_norm": 4.0, + "mode": "train", + "normalize_input": true, + "normalize_returns": true, + "num_batches_per_epoch": 2, + "num_envs_per_worker": 4, + "num_epochs": 2, + "num_policies": 1, + "num_workers": 32, + "obs_fps": 8.75, + "ordering_policy": "latest_ready", + "output_dir": "/home/ubuntu/lzj/mean-profiling/deadly_corridor/teacher-log", + "ppo_clip_ratio": 0.1, + "ppo_clip_value": 0.2, + "recurrence": 128, + "res_h": 72, + "res_w": 128, + "restart_behavior": "resume", + "rnn_size": 512, + "rnn_type": "gru", + "rollout": 128, + "save_every_sec": 600, + "seed": 42, + "simulated_worker_capacity": 1, + "simulator": "cpu", + "stats_avg": 100, + "train_dir": "/home/ubuntu/lzj/mean-profiling/deadly_corridor/teacher", + "train_for_env_steps": 25000000, + "use_rnn": true, + "value_loss_coeff": 0.5, + "wandb_group": "mean-profiling-3090-blackwell", + "wandb_job_type": "deadly_corridor_decision_history_native35_temporal", + "wandb_project": "starvla_tasks", + "wandb_tags": [ + "deadly_corridor_decision_history_native35_temporal_25m_seed42", + "deadly-corridor", + "frame_stack", + "temporal", + "qwenoft_rtx3090_temporal_hist8_capacity1_native35", + "fs4", + "seed42" + ], + "wandb_unique_id": "mean3090-deadly_corridor-teacher-blackwell-20260930-nice0", + "wide_aspect_ratio": false, + "with_wandb": true, + "worker_num_splits": 2 + }, + "command_line": "--mode train --algo APPO --env doom_deadly_corridor --experiment deadly_corridor_frame_stack_qwenoft_rtx3090_temporal_hist8_capacity1_native35_fs4_seed42 --train_dir results/checkpoints/deadly_corridor/decision_history_native35_temporal_25m_seed42 --restart_behavior resume --device gpu --seed 42 --episode_metrics_path results/checkpoints/deadly_corridor/decision_history_native35_temporal_25m_seed42/deadly_corridor_frame_stack_qwenoft_rtx3090_temporal_hist8_capacity1_native35_fs4_seed42/episode_metrics.jsonl --train_for_env_steps 25000000 --num_workers 32 --num_envs_per_worker 4 --num_policies 1 --batch_size 1024 --rollout 128 --recurrence 128 --num_epochs 2 --num_batches_per_epoch 2 --worker_num_splits 2 --learning_rate 0.0001 --gamma 0.99 --gae_lambda 0.95 --ppo_clip_ratio 0.1 --ppo_clip_value 0.2 --exploration_loss symmetric_kl --exploration_loss_coeff 0.001 --value_loss_coeff 0.5 --max_grad_norm 4.0 --rnn_type gru --rnn_size 512 --save_every_sec 600 --keep_checkpoints 5 --stats_avg 100 --experiment_summaries_interval 1 --async_rl True --use_rnn True --normalize_returns True --normalize_input True --latency-type profile_sample --latency-seed 42 --latency-profile-path results/profiling/profiles/openvla/1x-rtx3090/deadly_corridor/instance_a5037b165aa0cedc/profile.json --latency-profile-worker-slot 0 --simulated-worker-capacity 1 --add-latency-info False --eval-episodes 100 --eval-parallel-envs 1 --eval-max-steps 3600 --eval-deterministic True --eval-raw-reward --with_wandb True --wandb_project starvla_tasks --wandb_group deadly-corridor-fs4-qwenoft_rtx3090_temporal_hist8_capacity1_native35 --wandb_job_type deadly_corridor_decision_history_native35_temporal --wandb_tags deadly_corridor_decision_history_native35_temporal_25m_seed42 deadly-corridor frame_stack temporal qwenoft_rtx3090_temporal_hist8_capacity1_native35 fs4 seed42 --env-fps 35 --obs-fps 8.75 --res_w 128 --res_h 72 --wide_aspect_ratio False --frame_stack 4 --action-history-decisions 8 --simulator cpu --disable-reward-scaling True --hold-policy hold --ordering-policy latest_ready", + "continuous_tanh_scale": 0.0, + "decoder_mlp_layers": [], + "decorrelate_envs_on_one_worker": true, + "decorrelate_experience_max_seconds": 0, + "default_niceness": 0, + "device": "gpu", + "disable_reward_scaling": true, + "encoder_conv_architecture": "convnet_simple", + "encoder_conv_mlp_layers": [ + 512 + ], + "encoder_mlp_layers": [ + 512, + 512 + ], + "env": "doom_deadly_corridor", + "env_fps": 35.0, + "env_frameskip": 1, + "env_framestack": 1, + "env_gpu_actions": false, + "env_gpu_observations": true, + "episode_counter": false, + "episode_metrics_path": "/home/ubuntu/lzj/mean-profiling/deadly_corridor/teacher/deadly_corridor_mean_3090_s42/episode_metrics.jsonl", + "eval_deterministic": true, + "eval_env_frameskip": 1, + "eval_episodes": 100, + "eval_latency_raw_frame_values": "0,1,2,3,4,5,10,15,40", + "eval_max_steps": 3600, + "eval_parallel_envs": 1, + "eval_raw_reward": true, + "experiment": "deadly_corridor_mean_3090_s42", + "experiment_summaries_interval": 1, + "exploration_loss": "symmetric_kl", + "exploration_loss_coeff": 0.001, + "fasttd3_action_chunk_horizon": 1, + "fasttd3_actor_action_l2": 0.0, + "fasttd3_compile": true, + "fasttd3_replay_batch_size": 32768, + "fasttd3_replay_capacity": 6553600, + "fasttd3_train_for_optimizer_steps": 10000000000, + "fasttd3_transitions_per_update": 64, + "fasttd3_v_max": 250.0, + "fasttd3_v_min": -250.0, + "fixed_latency_ms": 73.69250777493353, + "flush_summaries_interval": 30, + "force_envs_single_thread": false, + "fps": 35, + "frame_stack": 4, + "gae_lambda": 0.95, + "gamma": 0.99, + "git_hash": "unknown", + "git_repo_name": "not a git repository", + "heartbeat_interval": 20, + "heartbeat_reporting_interval": 600, + "help": false, + "hold_policy": "hold", + "initial_model_path": null, + "initial_stddev": 1.0, + "keep_checkpoints": 5, + "kl_loss_coeff": 0.0, + "latency_profile_path": null, + "latency_profile_worker_slot": null, + "latency_seed": 42, + "latency_type": "fixed", + "learning_rate": 0.0001, + "load_checkpoint_kind": "latest", + "log_to_file": true, + "lr_adaptive_max": 0.01, + "lr_adaptive_min": 1e-06, + "lr_schedule": "constant", + "lr_schedule_kl_threshold": 0.008, + "max_episode_steps": 0, + "max_grad_norm": 4.0, + "max_latency_ms": null, + "max_pending_actions": null, + "max_policy_lag": 1000, + "mean_latency_ms": null, + "min_latency_ms": null, + "mode": "train", + "nonlinearity": "elu", + "normalize_input": true, + "normalize_input_keys": null, + "normalize_returns": true, + "num_agents": -1, + "num_batches_per_epoch": 2, + "num_batches_to_accumulate": 2, + "num_bots": -1, + "num_envs_per_worker": 4, + "num_epochs": 2, + "num_humans": 0, + "num_policies": 1, + "num_workers": 32, + "obs_fps": 8.75, + "obs_scale": 255.0, + "obs_subtract_mean": 0.0, + "optimizer": "adam", + "ordering_policy": "latest_ready", + "output_dir": "/home/ubuntu/lzj/mean-profiling/deadly_corridor/teacher-log", + "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", + "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, + "record_to": null, + "recurrence": 128, + "res_h": 72, + "res_w": 128, + "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, + "seed": 42, + "serial_mode": false, + "set_workers_cpu_affinity": true, + "shuffle_minibatches": false, + "simulated_worker_capacity": 1, + "simulator": "cpu", + "start_bot_difficulty": null, + "stats_avg": 100, + "std_latency_ms": null, + "summaries_use_frameskip": true, + "timelimit": null, + "train_dir": "/home/ubuntu/lzj/mean-profiling/deadly_corridor/teacher", + "train_for_env_steps": 25000000, + "train_for_seconds": 10000000000, + "use_env_info_cache": false, + "use_record_episode_statistics": false, + "use_rnn": true, + "value_bootstrap": false, + "value_loss_coeff": 0.5, + "vizdoom_init_lock_max_wait_sec": 60.0, + "vizdoom_init_lock_timeout_sec": 5.0, + "vizdoom_init_max_parallel": 10, + "vizdoom_init_retry_jitter_sec": 0.25, + "vtrace_c": 1.0, + "vtrace_rho": 1.0, + "wandb_dir": "/home/ubuntu/lzj/mean-profiling/deadly_corridor/wandb", + "wandb_group": "mean-profiling-3090-blackwell", + "wandb_job_type": "deadly_corridor_decision_history_native35_temporal", + "wandb_project": "starvla_tasks", + "wandb_tags": [ + "deadly_corridor_decision_history_native35_temporal_25m_seed42", + "deadly-corridor", + "frame_stack", + "temporal", + "qwenoft_rtx3090_temporal_hist8_capacity1_native35", + "fs4", + "seed42" + ], + "wandb_unique_id": "mean3090-deadly_corridor-teacher-blackwell-20260930-nice0", + "wandb_user": "dongqianyu99-zhejiang-university", + "wide_aspect_ratio": false, + "with_pbt": false, + "with_vtrace": false, + "with_wandb": true, + "worker_num_splits": 2 + }, + "checkpoint_experiment_dir": "/home/ubuntu/lzj/mean-profiling/deadly_corridor/teacher-publication", + "checkpoint_path": "/home/ubuntu/lzj/mean-profiling/deadly_corridor/teacher-publication/checkpoint.pth", + "checkpoint_train_step": 43436, + "config_source": "/home/ubuntu/lzj/mean-profiling/deadly_corridor/teacher-publication/config.json", + "context_image_count": 0, + "context_window": 1, + "env_fps": 35.0, + "env_name": "deadly_corridor", + "episode_return_field": "episode_raw_return", + "filter_preset": "none", + "filter_report": { + "accepted_episodes": 8500, + "accepted_train_episodes": 7650, + "accepted_val_episodes": 850, + "attempted_episodes": 8500, + "preset": "none", + "rejected_episodes": 0, + "train_rows_after_row_filter": 313925, + "train_rows_before_row_filter": 313925, + "val_rows_after_row_filter": 34465, + "val_rows_before_row_filter": 34465 + }, + "frame_stack": 4, + "image_resolution": [ + 120, + 160, + 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": 8500, + "max_steps_per_episode": 3600, + "max_steps_unit": "raw_frame", + "num_envs": 16, + "obs_fps": 8.75, + "obs_stride_raw_frames": 4, + "prompt_source": "sample_factory_checkpoint:base_prompt", + "replay_num_envs": 8, + "reward_field": "raw_reward", + "reward_semantics": "raw_reward", + "rows_unit": "decision_step", + "schema_version": 2, + "split_seed": 0, + "total_episodes": 8500, + "total_rows": 348390, + "train_episodes": 7650, + "train_rows": 313925, + "val_episodes": 850, + "val_fraction": 0.1, + "val_rows": 34465 +} diff --git a/latency-aware/deadly-corridor/mean_rtx3090_s42_8500ep_v1/demonstrations/raw/train.parquet b/latency-aware/deadly-corridor/mean_rtx3090_s42_8500ep_v1/demonstrations/raw/train.parquet new file mode 100644 index 0000000000000000000000000000000000000000..923366197bc506806c898efc036012a7961201bf --- /dev/null +++ b/latency-aware/deadly-corridor/mean_rtx3090_s42_8500ep_v1/demonstrations/raw/train.parquet @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3e3473cf7eeece0b3e107f516869b136159fd197ff4a8667a47e128cf336b9e2 +size 4950850571 diff --git a/latency-aware/deadly-corridor/mean_rtx3090_s42_8500ep_v1/demonstrations/raw/val.parquet b/latency-aware/deadly-corridor/mean_rtx3090_s42_8500ep_v1/demonstrations/raw/val.parquet new file mode 100644 index 0000000000000000000000000000000000000000..ad729f933b13f0d2740bd37bb87ad4bc000b18a4 --- /dev/null +++ b/latency-aware/deadly-corridor/mean_rtx3090_s42_8500ep_v1/demonstrations/raw/val.parquet @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:68456699331b2362acf90f859b9d5b379e844a6d38d6047c9f67d2cceb79b269 +size 543289255 diff --git a/latency-aware/deadly-corridor/mean_rtx3090_s42_8500ep_v1/demonstrations/validation/collection.complete.json b/latency-aware/deadly-corridor/mean_rtx3090_s42_8500ep_v1/demonstrations/validation/collection.complete.json new file mode 100644 index 0000000000000000000000000000000000000000..c24cb999f5efd62f56730771c945864bbef42488 --- /dev/null +++ b/latency-aware/deadly-corridor/mean_rtx3090_s42_8500ep_v1/demonstrations/validation/collection.complete.json @@ -0,0 +1,14 @@ +{ + "task": "deadly_corridor", + "teacher_checkpoint": "/home/ubuntu/lzj/mean-profiling/deadly_corridor/teacher-publication/checkpoint.pth", + "teacher_sha256": "e631e9017634f799860c581761b5fa48601109393c11d20a35187c6f75d14b88", + "episodes": 8500, + "train_episodes": 7650, + "val_episodes": 850, + "train_rows": 313925, + "val_rows": 34465, + "fixed_latency_ms": 73.69250777493353, + "action_horizon": 1, + "unique_seeds": 8500, + "state_dim": 120 +} diff --git a/latency-aware/deadly-corridor/mean_rtx3090_s42_8500ep_v1/demonstrations/validation/dataset.audit.json b/latency-aware/deadly-corridor/mean_rtx3090_s42_8500ep_v1/demonstrations/validation/dataset.audit.json new file mode 100644 index 0000000000000000000000000000000000000000..ef32f78d0ba24e883c57ed997aff0608dec2b61e --- /dev/null +++ b/latency-aware/deadly-corridor/mean_rtx3090_s42_8500ep_v1/demonstrations/validation/dataset.audit.json @@ -0,0 +1,36 @@ +{ + "task": "deadly_corridor", + "episodes": 8500, + "train_episodes": 7650, + "val_episodes": 850, + "unique_seeds": 8500, + "seed_range": [ + 42, + 8541 + ], + "split_intersection": 0, + "train_rows": 313925, + "val_rows": 34465, + "rows": 348390, + "state_dim": 120, + "history_decisions": 8, + "state_binary": true, + "action_dim": 7, + "source_joint_actions": 54, + "converted_native_binary_actions": true, + "finite_numeric": true, + "image_shape": [ + 120, + 160, + 3 + ], + "decoded_image_samples": 6, + "mean_ms": 73.69250777493353, + "effective_latency_raw_frames": 3, + "effective_latency_ms": 85.71428571428571, + "env_fps": 35, + "obs_fps": 8.75, + "raw_max_steps": 3600, + "raw_parquet_bytes": 5494139826, + "converted_all_episode_row_counts_verified": true +} diff --git a/latency-aware/deadly-corridor/mean_rtx3090_s42_8500ep_v1/evidence/collect_sf.py b/latency-aware/deadly-corridor/mean_rtx3090_s42_8500ep_v1/evidence/collect_sf.py new file mode 100644 index 0000000000000000000000000000000000000000..3ccf71f870cc1c2479dfcde9ba04553e0af49b21 --- /dev/null +++ b/latency-aware/deadly-corridor/mean_rtx3090_s42_8500ep_v1/evidence/collect_sf.py @@ -0,0 +1,18 @@ +from pathlib import Path +import argparse,hashlib,json,subprocess,sys +import pyarrow.parquet as pq +from latency_bench.data.exporters.starvla_lerobot import convert_canonical_rollout_to_lerobot +ROOT=Path('/home/ubuntu/lzj/mean-profiling');REPO=Path('/home/ubuntu/lzj/latency-sensitive-bench') +BUDGETS={'flappy':(200,18000,256,256,'gpu'),'deadly_corridor':(8500,3600,16,8,'cpu'),'ant':(100,1000,12,8,'gpu')} +p=argparse.ArgumentParser();p.add_argument('task',choices=BUDGETS);p.add_argument('--finalize-only',action='store_true');args=p.parse_args() +task=args.task;out=ROOT/task;cfg=json.loads((out/'teacher.flat.json').read_text());ck=Path(json.loads((out/'teacher.upload.json').read_text())['local_checkpoint']);episodes,cap,probe,replay,device=BUDGETS[task];raw=out/'data/raw' +command=[sys.executable,str(REPO/'scripts/rollout_data/export_sf_teacher_dataset.py'),'--checkpoint-root',str(ck),'--output-dir',str(raw),'--max-episodes',str(episodes),'--max-steps-per-episode',str(cap),'--num-envs',str(probe),'--replay-num-envs',str(replay),'--device',device,'--val-fraction','0.1','--split-seed','0','--context-window','1','--filter-preset','none','--deterministic','--image-writer-workers','32' if device=='gpu' else '8'] +if task=='ant':command.extend(['--episode-return-gt','2500','--export-image-size','224,224']) +else:command.extend(['--max-attempts',str(episodes)]) +(out/'collection-command.json').write_text(json.dumps(command,indent=2)+'\n') +if not args.finalize_only: + subprocess.run(command,cwd=REPO,check=True) + convert_canonical_rollout_to_lerobot(raw,out/'data/lerobot',action_carrier='native') +meta=json.loads((raw/'metadata.json').read_text());manifest=json.loads((out/'data/lerobot/manifest.json').read_text());assert meta['total_episodes']==episodes and manifest['state_dim']=={'flappy':24,'deadly_corridor':120,'ant':27}[task] +train=set(pq.read_table(raw/'train.parquet',columns=['seed']).column('seed').unique().to_pylist());val=set(pq.read_table(raw/'val.parquet',columns=['seed']).column('seed').unique().to_pylist());assert len(train)==meta['train_episodes'] and len(val)==meta['val_episodes'] and train.isdisjoint(val) +report={'task':task,'teacher_checkpoint':str(ck),'teacher_sha256':hashlib.sha256(ck.read_bytes()).hexdigest(),'episodes':episodes,'train_episodes':meta['train_episodes'],'val_episodes':meta['val_episodes'],'train_rows':meta['train_rows'],'val_rows':meta['val_rows'],'fixed_latency_ms':cfg['fixed_latency_ms'],'action_horizon':1,'unique_seeds':len(train|val),'state_dim':manifest['state_dim']};(out/'data/collection.complete.json').write_text(json.dumps(report,indent=2)+'\n');print(json.dumps(report),flush=True) diff --git a/latency-aware/deadly-corridor/mean_rtx3090_s42_8500ep_v1/evidence/collection-command.json b/latency-aware/deadly-corridor/mean_rtx3090_s42_8500ep_v1/evidence/collection-command.json new file mode 100644 index 0000000000000000000000000000000000000000..b8574270b83178c0c8c8b3416b6bb10645e19a5e --- /dev/null +++ b/latency-aware/deadly-corridor/mean_rtx3090_s42_8500ep_v1/evidence/collection-command.json @@ -0,0 +1,31 @@ +[ + "/home/ubuntu/lzj/conda/envs/sf-data/bin/python", + "/home/ubuntu/lzj/latency-sensitive-bench/scripts/rollout_data/export_sf_teacher_dataset.py", + "--checkpoint-root", + "/home/ubuntu/lzj/mean-profiling/deadly_corridor/teacher-publication/checkpoint.pth", + "--output-dir", + "/home/ubuntu/lzj/mean-profiling/deadly_corridor/data/raw", + "--max-episodes", + "8500", + "--max-steps-per-episode", + "3600", + "--num-envs", + "16", + "--replay-num-envs", + "8", + "--device", + "cpu", + "--val-fraction", + "0.1", + "--split-seed", + "0", + "--context-window", + "1", + "--filter-preset", + "none", + "--deterministic", + "--image-writer-workers", + "8", + "--max-attempts", + "8500" +] diff --git a/latency-aware/deadly-corridor/mean_rtx3090_s42_8500ep_v1/evidence/doom-mean-reset.patch b/latency-aware/deadly-corridor/mean_rtx3090_s42_8500ep_v1/evidence/doom-mean-reset.patch new file mode 100644 index 0000000000000000000000000000000000000000..9a9870a848a8ab9d6b5026b4cc92d80fde56dabe --- /dev/null +++ b/latency-aware/deadly-corridor/mean_rtx3090_s42_8500ep_v1/evidence/doom-mean-reset.patch @@ -0,0 +1,192 @@ +diff --git a/latency_bench/data/sf_teacher_rollout.py b/latency_bench/data/sf_teacher_rollout.py +index ef45d95b..e32b942e 100644 +--- a/latency_bench/data/sf_teacher_rollout.py ++++ b/latency_bench/data/sf_teacher_rollout.py +@@ -964,6 +964,11 @@ def _replay_accepted_specs( + row_buffer_size=int(row_buffer_size), + export_image_size=export_image_size, + ) ++ stream_current_images = ( ++ image_transform_config.normalized_transform() == "raw_rgb" ++ and int(context_window) <= 1 ++ and filter_preset == "none" ++ ) + image_resolution: list[int] | None = None + + def assign_specs_to_slots(slots: list[int]) -> dict[int, int]: +@@ -1051,7 +1056,10 @@ def _replay_accepted_specs( + next_raw_frame=next_decision_step * obs_stride_raw_frames, + max_steps_per_episode=int(max_steps_per_episode), + ) +- active_episode_frames[slot].append(np.array(image_frame, copy=True)) ++ if stream_current_images: ++ image_writer.write(image_frame, image_path) ++ else: ++ active_episode_frames[slot].append(np.array(image_frame, copy=True)) + row = teacher_dataset_row( + env_name=env_name, + spec=spec, +@@ -1104,7 +1112,9 @@ def _replay_accepted_specs( + ) + else: + shard_rows = [dict(row) for row in final_rows] +- if _is_demon_ghost_trail: ++ if stream_current_images: ++ image_writer.flush() ++ elif _is_demon_ghost_trail: + # Frames are pre-built ghost trail images; map directly. + ep_steps = [int(row["decision_step"]) for row in final_rows] + frames_by_decision_step = { +@@ -1118,12 +1128,13 @@ def _replay_accepted_specs( + frames=active_episode_frames[slot], + config=image_transform_config, + ) +- shard_rows = _write_filtered_episode_images( +- image_writer, +- shard_rows, +- frames_by_decision_step, +- context_window=int(context_window), +- ) ++ if not stream_current_images: ++ shard_rows = _write_filtered_episode_images( ++ image_writer, ++ shard_rows, ++ frames_by_decision_step, ++ context_window=int(context_window), ++ ) + write_episode_row_shard(output_dir, shard_rows) + progress.update(1) + replayed += 1 +@@ -1272,9 +1283,8 @@ def _build_rollout_driver_runtime( + driver = BatchedTeacherRolloutDriver(runtime["envs"][0], int(num_envs)) + else: + recreate_env = None +- if integration_name == "deadly_corridor" and ( +- "export_zero_action_history" in config or getattr(cfg, "rollout_profile_seed_by_reset", False) +- ): ++ # ViZDoom applies reset seeds only while creating its native game. ++ if integration_name == "deadly_corridor": + from sample_factory.algo.utils.make_env import make_env_func_batched + from sample_factory.utils.attr_dict import AttrDict + +diff --git a/tests/data/test_sf_teacher_dataset_export.py b/tests/data/test_sf_teacher_dataset_export.py +index e6b25233..61a58118 100644 +--- a/tests/data/test_sf_teacher_dataset_export.py ++++ b/tests/data/test_sf_teacher_dataset_export.py +@@ -498,8 +498,11 @@ def test_serial_rollout_driver_continues_life_loss_with_seedless_reset(monkeypat + assert info == {"reset_seed": None} + + +-def test_serial_teacher_replay_starts_each_seed_in_a_fresh_environment(monkeypatch): +- sf_teacher_drivers = _load_sf_teacher_drivers_for_direct_test(monkeypatch) ++@pytest.mark.parametrize("latency_type", ["fixed", "profile_sample"]) ++def test_serial_teacher_replay_starts_each_seed_in_a_fresh_environment(monkeypatch, tmp_path, latency_type): ++ from contextlib import nullcontext ++ from latency_bench.data import sf_teacher_rollout ++ from sample_factory.algo.utils import make_env + + class StatefulEnv: + action_space = object() +@@ -514,13 +517,17 @@ def test_serial_teacher_replay_starts_each_seed_in_a_fresh_environment(monkeypat + def close(self): + return None + +- driver = sf_teacher_drivers.SerialTeacherRolloutDriver( +- [StatefulEnv()], recreate_env=lambda _slot: StatefulEnv() ++ cfg = SimpleNamespace(frame_stack=4, action_history_decisions=8, latency_type=latency_type) ++ monkeypatch.setattr(sf_teacher_rollout, "sample_factory_init_log", lambda _: nullcontext()) ++ monkeypatch.setattr(sf_teacher_rollout, "build_sample_factory_rollout_cfg", lambda _: (object(), cfg)) ++ monkeypatch.setattr(sf_teacher_rollout, "build_env_and_policy", lambda *args: {"envs": [StatefulEnv()]}) ++ monkeypatch.setattr(make_env, "make_env_func_batched", lambda *args, **kwargs: StatefulEnv()) ++ _, _, driver, _ = sf_teacher_rollout._build_rollout_driver_runtime( ++ {"env_name": "deadly_corridor", "integration_name": "deadly_corridor"}, ++ output_dir=tmp_path, num_envs=1, export_env_raw_rgb_frames=False, + ) +- + driver.reset_slots({0: 42}) + observation, _ = driver.reset_slots({0: 43})[0] +- + assert observation["obs"].tolist() == [43, 1] + + +@@ -670,19 +677,22 @@ def test_collect_teacher_dataset_probe_retries_until_episode_is_accepted(monkeyp + def action_distribution(self): + return object() + ++ reset_seeds = [] ++ + class FakeEnv: + def __init__(self): + self.action_space = object() +- self.reset_seeds: list[int] = [] ++ self.reset_seeds = reset_seeds + + def reset(self, seed=None): + self.reset_seeds.append(seed) ++ self.current_seed = seed + return np.zeros((3, 2, 2), dtype=np.uint8), [ + {"env_raw_rgb_frame_stack": _env_raw_rgb_frame_stack(5)} + ] + + def step(self, action): +- reward = 3000.0 if len(self.reset_seeds) == 1 else 3000.001 ++ reward = 3000.0 if self.current_seed == 11 else 3000.001 + return ( + np.zeros((3, 2, 2), dtype=np.uint8), + np.float32(reward), +@@ -702,6 +712,8 @@ def test_collect_teacher_dataset_probe_retries_until_episode_is_accepted(monkeyp + + fake_env = FakeEnv() + fake_actor = FakeActorCritic() ++ from sample_factory.algo.utils import make_env ++ monkeypatch.setattr(make_env, "make_env_func_batched", lambda *args, **kwargs: FakeEnv()) + fake_cfg = SimpleNamespace(frame_stack=1, env_fps=35.0, obs_fps=17.5, seed=11) + + monkeypatch.setattr(exporter, "build_sample_factory_rollout_cfg", lambda config: (object(), fake_cfg)) +@@ -1273,7 +1285,8 @@ def test_collect_teacher_dataset_saves_demon_attack_latest_raw_rgb_image(monkeyp + + + @pytest.mark.parametrize("export_history", [False, True]) +-def test_collect_teacher_dataset_saves_flappy_latest_raw_rgb_image(monkeypatch, tmp_path, export_history): ++@pytest.mark.parametrize("filter_preset", ["clean_v1", "none"]) ++def test_collect_teacher_dataset_saves_flappy_latest_raw_rgb_image(monkeypatch, tmp_path, export_history, filter_preset): + import torch + from gymnasium.spaces import Discrete + from PIL import Image +@@ -1307,6 +1320,9 @@ def test_collect_teacher_dataset_saves_flappy_latest_raw_rgb_image(monkeypatch, + class FakeEnv: + action_space = Discrete(2) + ++ def __init__(self, exporting): ++ self.exporting = exporting ++ + def reset(self, seed=None): + self.steps = 0 + obs = {"obs": np.zeros((1, 12, 2, 2), dtype=np.uint8)} +@@ -1323,6 +1339,9 @@ def test_collect_teacher_dataset_saves_flappy_latest_raw_rgb_image(monkeypatch, + + def step(self, action): + self.steps += 1 ++ if self.exporting and self.steps == 2 and filter_preset == "none": ++ image_path = output_dir / "images/train/episode_000000/decision_step_000000.png" ++ assert image_path.is_file(), "First frame must be on disk before the episode ends" + obs = {"obs": np.ones((1, 12, 2, 2), dtype=np.uint8)} + next_stack = np.stack([np.full((6, 5, 3), 50 + idx, dtype=np.uint8) for idx in range(4)], axis=0) + return ( +@@ -1354,7 +1373,7 @@ def test_collect_teacher_dataset_saves_flappy_latest_raw_rgb_image(monkeypatch, + exporter, + "build_env_and_policy", + lambda module, cfg, config, num_envs: { +- "envs": [FakeEnv()], ++ "envs": [FakeEnv(cfg.export_env_raw_rgb_frames)], + "env_info": object(), + "actor_critic": FakeActorCritic(), + "device": torch.device("cpu"), +@@ -1382,6 +1401,7 @@ def test_collect_teacher_dataset_saves_flappy_latest_raw_rgb_image(monkeypatch, + export_image_size=None, + image_writer_workers=0, + row_buffer_size=1, ++ filter_preset=filter_preset, + ) + train_rows = _read_shard_rows(output_dir, "train") + val_rows = _read_shard_rows(output_dir, "val") diff --git a/latency-aware/deadly-corridor/mean_rtx3090_s42_8500ep_v1/evidence/profile.source.json b/latency-aware/deadly-corridor/mean_rtx3090_s42_8500ep_v1/evidence/profile.source.json new file mode 100644 index 0000000000000000000000000000000000000000..333844eb05ec67a5791a289ee39c227c95991641 --- /dev/null +++ b/latency-aware/deadly-corridor/mean_rtx3090_s42_8500ep_v1/evidence/profile.source.json @@ -0,0 +1,66 @@ +{ + "burst_model_path": "latency_burst_model.json", + "distribution_path": "latency_distribution.json", + "env_fps": 60, + "frame_ms": 16.666666666666668, + "gpu_class": "1x-rtx3090", + "instance_id": "instance_a5037b165aa0cedc", + "latency_kind": "observation_to_action_latency", + "latency_method": "temporal", + "model_id": "openvla", + "n_admitted_observations": 506, + "n_capacity_drops": 493, + "n_observation_attempts": 999, + "per_slot_summary": { + "0": { + "admitted_count": 506, + "mean_observation_to_action_latency_ms": 73.69250777493353, + "mean_worker_service_time_ms": 73.00619814047229, + "p95_observation_to_action_latency_ms": 78.0192643404007, + "p95_worker_service_time_ms": 76.60398542881012, + "p99_worker_service_time_ms": 93.66708476543427 + } + }, + "provenance": { + "base_config": "configs/experiments/deadly_corridor/starvla_profile.yaml", + "checkpoint_kind": "best", + "model_artifact": { + "checkpoint": "checkpoints/steps_500_state/model.safetensors", + "model_config": "config.full.yaml", + "path_in_repo": ".", + "repo_id": "talha15032/openvla_bridge_deadly_corridor_single_latency_clean_data_bce_exp2", + "source": "local" + }, + "session_ids": [ + 0, + 1, + 2, + 3, + 4 + ] + }, + "sample_model_type": "hidden_regime", + "source_run_id": "20260914T171446047509Z", + "summary": { + "frame_ms": 16.666666666666668, + "max_ms": 101.71253681182861, + "mean_effective_frames": 4.4215504664960115, + "mean_ms": 73.69250777493353, + "min_ms": 68.33836698532104, + "n_samples": 506, + "p50_frames": 4.3794349193572994, + "p50_ms": 72.99058198928833, + "p90_frames": 4.606898102760314, + "p90_ms": 76.78163504600525, + "p95_frames": 4.681155860424042, + "p95_ms": 78.0192643404007, + "p99_frames": 5.741582728385924, + "p99_ms": 95.69304547309875, + "prob_latency_gt_1_frame": 1.0, + "prob_latency_gt_2_frames": 1.0, + "prob_latency_gt_3_frames": 1.0, + "std_ms": 4.766097168829327 + }, + "visualization_path": "latency_profile.png", + "workload_id": "deadly_corridor" +} diff --git a/latency-aware/deadly-corridor/mean_rtx3090_s42_8500ep_v1/evidence/starvla_lerobot.source.py b/latency-aware/deadly-corridor/mean_rtx3090_s42_8500ep_v1/evidence/starvla_lerobot.source.py new file mode 100644 index 0000000000000000000000000000000000000000..fba03402d2c26da376833bead62f0b17d4fed9cd --- /dev/null +++ b/latency-aware/deadly-corridor/mean_rtx3090_s42_8500ep_v1/evidence/starvla_lerobot.source.py @@ -0,0 +1,1141 @@ +"""Convert canonical rollout exports to the StarVLA LeRobot image format.""" + +from __future__ import annotations + +import io +import json +import shutil +from pathlib import Path +from typing import Any, Mapping + +import numpy as np +import pyarrow as pa +import pyarrow.parquet as pq +from PIL import Image + +from latency_bench.utils.io import write_jsonl + +from latency_bench.data.action_layouts import ( + ACTION_LAYOUT_GYMNASIUM_CONTINUOUS_V1, + ACTION_LAYOUT_GYMNASIUM_DISCRETE_V1, + ACTION_LAYOUT_G1_SONIC_ABSOLUTE_JOINT_TARGET_V1, + action_spec_for_env, +) +from latency_bench.data.filters import density_weights_clean_v1 +from latency_bench.data.state_normalization import min_max_normalize_state + +STATE_DIM = 1 +BRIDGE_ACTION_DIM = 7 +LE_ROBOT_CHUNK_SIZE = 1000 +_ROBOT_TYPES_BY_ENV = { + "flappy": "rl_games_flappy", + "demon_attack": "rl_games_demon_attack", + "deadly_corridor": "rl_games_deadly_corridor", + "atlantis": "rl_games_atlantis", + "asterix": "rl_games_asterix", +} +_HISTORY_STATE_DIMS = {"flappy": 24, "demon_attack": 88, "deadly_corridor": 120, "atlantis": 56, "asterix": 136} +_DEADLY_MULTIBINARY_LABELS = [ + "MOVE_FORWARD", "MOVE_BACKWARD", "MOVE_LEFT", "MOVE_RIGHT", + "TURN_LEFT", "TURN_RIGHT", "ATTACK", +] + + +def write_starvla_modality( + dataset_dir: Path | str, + state_dim: int, + action_dim: int, + *, + action_key: str = "action", +) -> Path: + path = Path(dataset_dir) / "meta" / "modality.json" + path.parent.mkdir(parents=True, exist_ok=True) + path.write_text( + json.dumps( + { + "state": { + "state": { + "start": 0, + "end": state_dim, + "dtype": "float32", + "absolute": True, + "original_key": "observation.state", + } + }, + "action": { + action_key: { + "start": 0, + "end": action_dim, + "dtype": "float32", + "absolute": True, + "original_key": action_key, + } + }, + "video": {"ego_view": {"original_key": "observation.image"}}, + "annotation": { + "human.task_description": {"original_key": "task_index"} + }, + }, + indent=2, + ) + + "\n", + encoding="utf-8", + ) + return path + + +def _action_carrier_spec( + action_labels: list[str], + action_carrier: str, +) -> tuple[list[str], int, int, list[str]]: + if action_carrier == "native": + return action_labels, len(action_labels), STATE_DIM, ["state"] + if action_carrier != "bridge": + raise ValueError(f"Unsupported action_carrier={action_carrier!r}") + if len(action_labels) > BRIDGE_ACTION_DIM: + raise ValueError( + f"bridge action carrier supports at most {BRIDGE_ACTION_DIM} actions, " + f"got {len(action_labels)}" + ) + carrier_labels = [ + *action_labels, + *( + f"BRIDGE_PAD_{index}" + for index in range(len(action_labels), BRIDGE_ACTION_DIM) + ), + ] + state_labels = [f"BRIDGE_STATE_{index}" for index in range(BRIDGE_ACTION_DIM)] + return carrier_labels, len(action_labels), BRIDGE_ACTION_DIM, state_labels + + +def _robot_type_for_env(env_name: str, integration_name: str) -> str: + if integration_name == "gymnasium": + return "rl_games_gymnasium" + return _ROBOT_TYPES_BY_ENV[env_name] + + +def _write_starvla_handoff_files( + output_dir: Path, + val_output_dir: Path, + *, + env_name: str, + integration_name: str, + rows: list[dict[str, Any]], +) -> tuple[Path, Path]: + prompt_map = { + str(int(row["latency_raw_frames"])): { + "prompt": row["prompt"], + "latency_raw_frames": int(row["latency_raw_frames"]), + "latency_ms": float(row["latency_ms"]), + } + for row in rows + } + prompt_map_path = output_dir / "latency_prompt_map.json" + prompt_map_path.write_text(json.dumps(prompt_map, indent=2), encoding="utf-8") + + mixture_path = output_dir.parent / "_generated_mixtures" / f"{output_dir.name}.json" + mixture_path.parent.mkdir(parents=True, exist_ok=True) + robot_type = _robot_type_for_env(env_name, integration_name) + mixture_path.write_text( + json.dumps( + { + output_dir.name: [[output_dir.name, 1.0, robot_type]], + val_output_dir.name: [[val_output_dir.name, 1.0, robot_type]], + }, + indent=2, + ), + encoding="utf-8", + ) + return prompt_map_path, mixture_path + + +def _png_bytes(image: Any) -> bytes: + """Serialize a rollout image value to PNG bytes. + + Args: + image: Hugging Face image struct, PIL image, or array-like image. + + Returns: + RGB PNG bytes. + """ + if isinstance(image, Mapping): + if image["bytes"] is not None: + return bytes(image["bytes"]) + with Path(image["path"]).open("rb") as handle: + return handle.read() + if not isinstance(image, Image.Image): + image = Image.fromarray(np.asarray(image)) + image = image.convert("RGB") + buffer = io.BytesIO() + image.save(buffer, format="PNG") + return buffer.getvalue() + + +def _load_rows(raw_export_dir: Path, split: str) -> list[dict[str, Any]]: + """Load canonical rollout rows from one parquet split. + + Args: + raw_export_dir: Directory containing ``train.parquet`` and ``val.parquet``. + split: Split name to read. + + Returns: + Rows as Python dictionaries. + """ + path = raw_export_dir / f"{split}.parquet" + # Bound nested image buffers per Arrow batch; the converter still owns all rows. + return [ + row + for batch in pq.ParquetFile(path).iter_batches(batch_size=1024) + for row in batch.to_pylist() + ] + + +def _episode_groups(rows: list[dict[str, Any]]) -> list[list[dict[str, Any]]]: + """Group canonical rows by episode and order by decision step. + + Args: + rows: Canonical rollout rows from one split. + + Returns: + Episode row groups sorted by episode index. + """ + grouped: dict[int, list[dict[str, Any]]] = {} + for row in rows: + grouped.setdefault(int(row["episode_idx"]), []).append(row) + return [ + sorted(episode_rows, key=lambda row: int(row["decision_step"])) + for _, episode_rows in sorted(grouped.items()) + ] + + +def _legacy_row_action_vector(row: dict[str, Any], *, action_dim: int) -> list[float]: + if "action" in row and row["action"] is not None: + return [float(value) for value in row["action"]] + if "action_id" not in row: + raise KeyError(f"row is missing action fields; available columns: {sorted(row)}") + action_id = int(row["action_id"]) + if action_id < 0 or action_id >= action_dim: + raise ValueError( + f"action_id={action_id} is outside action range [0, {action_dim - 1}]" + ) + return [1.0 if index == action_id else 0.0 for index in range(action_dim)] + + +def _write_legacy_episode( + path: Path, + rows: list[dict[str, Any]], + *, + action_dim: int, +) -> None: + path.parent.mkdir(parents=True, exist_ok=True) + table = pa.table( + { + "observation.image": pa.array( + [{"bytes": row["image_bytes"], "path": None} for row in rows], + type=pa.struct([("bytes", pa.binary()), ("path", pa.string())]), + ), + "observation.state": pa.array( + [[0.0] * STATE_DIM for _ in rows], + type=pa.list_(pa.float32(), STATE_DIM), + ), + "action": pa.array( + [row["action"] for row in rows], + type=pa.list_(pa.float32(), action_dim), + ), + "timestamp": pa.array( + [row["timestamp"] for row in rows], type=pa.float64() + ), + "episode_index": pa.array( + [row["episode_index"] for row in rows], type=pa.int64() + ), + "frame_index": pa.array( + [row["frame_index"] for row in rows], type=pa.int64() + ), + "task_index": pa.array( + [row["task_index"] for row in rows], type=pa.int64() + ), + "done": pa.array([row["done"] for row in rows], type=pa.bool_()), + "reward": pa.array([row["reward"] for row in rows], type=pa.float32()), + "density_weight": pa.array( + [row.get("density_weight", 1.0) for row in rows], + type=pa.float32(), + ), + } + ) + pq.write_table(table, path) + + +def _write_legacy_metadata( + dataset_dir: Path, + *, + fps: float, + action_labels: list[str], + episode_lengths: list[int], + task_prompts: list[str], +) -> None: + meta_dir = dataset_dir / "meta" + meta_dir.mkdir(parents=True, exist_ok=True) + action_dim = len(action_labels) + (meta_dir / "info.json").write_text( + json.dumps( + { + "codebase_version": "v2.0", + "fps": fps, + "features": { + "observation.image": {"dtype": "image"}, + "observation.state": { + "dtype": "float32", + "shape": [STATE_DIM], + }, + "action": { + "dtype": "float32", + "shape": [action_dim], + "names": action_labels, + }, + "timestamp": {"dtype": "float64", "shape": [1]}, + "episode_index": {"dtype": "int64", "shape": [1]}, + "frame_index": {"dtype": "int64", "shape": [1]}, + "task_index": {"dtype": "int64", "shape": [1]}, + }, + }, + indent=2, + ), + encoding="utf-8", + ) + write_jsonl( + meta_dir / "episodes.jsonl", + [ + {"episode_index": index, "length": int(length)} + for index, length in enumerate(episode_lengths) + ], + ) + write_jsonl( + meta_dir / "tasks.jsonl", + [ + {"task_index": index, "task": prompt} + for index, prompt in enumerate(task_prompts) + ], + ) + + +def _convert_legacy_split( + rows: list[dict[str, Any]], + output_dir: Path, + *, + env_name: str, + action_labels: list[str], + fps: float, + reward_field: str, + density_weighting: bool, +) -> dict[str, Any]: + output_dir.mkdir(parents=True, exist_ok=True) + prompt_to_task_index: dict[str, int] = {} + task_prompts: list[str] = [] + episode_lengths: list[int] = [] + for new_episode_idx, episode_rows in enumerate(_episode_groups(rows)): + density_weights = ( + density_weights_clean_v1(env_name, episode_rows) + if density_weighting + else [1.0] * len(episode_rows) + ) + out_rows = [] + for frame_idx, row in enumerate(episode_rows): + row["latency_raw_frames"] + prompt = str(row["prompt"]) + if prompt not in prompt_to_task_index: + prompt_to_task_index[prompt] = len(task_prompts) + task_prompts.append(prompt) + out_rows.append( + { + "image_bytes": _png_bytes(row["image"]), + "action": _legacy_row_action_vector( + row, + action_dim=len(action_labels), + ), + "timestamp": frame_idx / fps, + "episode_index": new_episode_idx, + "frame_index": frame_idx, + "task_index": prompt_to_task_index[prompt], + "done": frame_idx == len(episode_rows) - 1, + "reward": float(row[reward_field]), + "density_weight": float(density_weights[frame_idx]), + } + ) + episode_lengths.append(len(out_rows)) + episode_chunk = new_episode_idx // 1000 + _write_legacy_episode( + output_dir + / f"data/chunk-{episode_chunk:03d}/episode_{new_episode_idx:06d}.parquet", + out_rows, + action_dim=len(action_labels), + ) + _write_legacy_metadata( + output_dir, + fps=fps, + action_labels=action_labels, + episode_lengths=episode_lengths, + task_prompts=task_prompts, + ) + return { + "dataset_name": output_dir.name, + "env_name": env_name, + "episodes": len(episode_lengths), + "frames": int(sum(episode_lengths)), + "task_prompts": task_prompts, + } + + +def _convert_legacy_export( + raw_export_dir: Path, + output_dir: Path, + *, + metadata: dict[str, Any], + force: bool, +) -> dict[str, Any]: + env_name = str(metadata["env_name"]) + action_spec = action_spec_for_env(env_name) + action_labels = list(metadata["action_labels"]) + reward_field = str(metadata["reward_field"]) + rows_unit = metadata["rows_unit"] + if rows_unit != "decision_step": + raise ValueError(f"rows_unit must be 'decision_step', got {rows_unit!r}") + fps = metadata["obs_fps"] + obs_stride = metadata["obs_stride_raw_frames"] + val_output_dir = output_dir.with_name(f"{output_dir.name}__val") + if output_dir.exists() and force: + shutil.rmtree(output_dir) + if val_output_dir.exists() and force: + shutil.rmtree(val_output_dir) + train_manifest = _convert_legacy_split( + _load_rows(raw_export_dir, "train"), + output_dir, + env_name=env_name, + action_labels=action_labels, + fps=fps, + reward_field=reward_field, + density_weighting=True, + ) + val_manifest = _convert_legacy_split( + _load_rows(raw_export_dir, "val"), + val_output_dir, + env_name=env_name, + action_labels=action_labels, + fps=fps, + reward_field=reward_field, + density_weighting=False, + ) + manifest = { + **train_manifest, + "format": "starvla_lerobot_v2_image_parquet", + "source": str(raw_export_dir), + "action_layout": str(metadata["action_layout"]), + "action_labels": action_labels, + "action_dim": len(action_labels), + "active_action_dim": len(action_spec.labels), + "rows_unit": rows_unit, + "obs_fps": fps, + "obs_stride_raw_frames": obs_stride, + "validation_dataset_name": val_output_dir.name, + "validation_episodes": val_manifest["episodes"], + "validation_frames": val_manifest["frames"], + } + (output_dir / "manifest.json").write_text( + json.dumps(manifest, indent=2), + encoding="utf-8", + ) + return manifest + + +def _convert_history_game_export( + raw_export_dir: Path, + output_dir: Path, + *, + metadata: dict[str, Any], + force: bool, +) -> dict[str, Any]: + """Convert decision-v1 game rollouts with their real pre-decision state.""" + env_name = metadata["env_name"] + state_dim = _HISTORY_STATE_DIMS[env_name] + action_labels = ( + _DEADLY_MULTIBINARY_LABELS + if env_name == "deadly_corridor" else list(metadata["action_labels"]) + ) + state_labels = [f"decision_history_{index}" for index in range(state_dim)] + train_rows = _load_rows(raw_export_dir, "train") + val_rows = _load_rows(raw_export_dir, "val") + for row in [*train_rows, *val_rows]: + if len(row["state"]) != state_dim: + raise ValueError(f"{env_name} decision history has {len(row['state'])} values, expected {state_dim}") + if env_name == "deadly_corridor": + action_id = row["action_id"] + turn, move, strafe, attack = ( + action_id // 18, (action_id // 6) % 3, (action_id // 2) % 3, action_id % 2 + ) + row["action"] = [ + float(move == 1), float(move == 2), float(strafe == 1), + float(strafe == 2), float(turn == 1), float(turn == 2), + float(attack == 1), + ] + first_row = (train_rows or val_rows)[0] + with Image.open(io.BytesIO(_png_bytes(first_row["image"]))) as image: + image_shape = (image.height, image.width, 3) + val_output_dir = output_dir.with_name(f"{output_dir.name}__val") + if output_dir.exists() and force: + shutil.rmtree(output_dir) + if val_output_dir.exists() and force: + shutil.rmtree(val_output_dir) + common = { + "env_name": env_name, + "integration_name": env_name, + "action_labels": action_labels, + "active_action_dim": len(action_labels), + "state_dim": state_dim, + "state_labels": state_labels, + "context_image_count": None, + "image_shape": image_shape, + "fps": metadata["obs_fps"], + "reward_field": metadata["reward_field"], + "state_normalization": None, + "include_state": True, + "density_weighting": False, + } + train_manifest = _convert_split(train_rows, output_dir, **common) + val_manifest = _convert_split(val_rows, val_output_dir, **common) + for directory in (output_dir, val_output_dir): + modality_path = directory / "meta/modality.json" + modality = json.loads(modality_path.read_text()) + modality["annotation"] = { + "human.action.task_description": {"original_key": "task_index"} + } + modality_path.write_text(json.dumps(modality, indent=2) + "\n") + prompt_map_path, mixture_path = _write_starvla_handoff_files( + output_dir, val_output_dir, env_name=env_name, + integration_name=env_name, rows=[*train_rows, *val_rows], + ) + contract = { + "format": "starvla_lerobot_v2_decision_v1", + "source": str(raw_export_dir), + "robot_type": _ROBOT_TYPES_BY_ENV[env_name], + "action_layout": "multibinary_7" if env_name == "deadly_corridor" else metadata["action_layout"], + "source_action_layout": metadata["action_layout"], + "action_labels": action_labels, + "action_dim": len(action_labels), + "active_action_dim": len(action_labels), + "state_source": "decision_v1", + "state_dim": state_dim, + "state_labels": state_labels, + "state_normalization": {"min": [0.0] * state_dim, "max": [1.0] * state_dim}, + "action_history_decisions": metadata["action_history_decisions"], + "rows_unit": metadata["rows_unit"], + "obs_fps": metadata["obs_fps"], + "obs_stride_raw_frames": metadata["obs_stride_raw_frames"], + "latency_prompt_map_path": str(prompt_map_path), + "custom_mixtures_path": str(mixture_path), + } + if env_name in {"atlantis", "asterix"}: + # The exported student image is current RGB; four-frame stacking belongs + # to the teacher and remains recorded in the raw collection metadata. + contract.update(integration_name="gymnasium", + gymnasium_task={**metadata["gymnasium_task"], "frame_stack": 1}) + manifest = {**train_manifest, **contract, + "validation_dataset_name": val_output_dir.name, + "validation_episodes": val_manifest["episodes"], + "validation_frames": val_manifest["frames"]} + (output_dir / "manifest.json").write_text(json.dumps(manifest, indent=2) + "\n") + (val_output_dir / "manifest.json").write_text( + json.dumps({**val_manifest, **contract, "training_dataset_name": output_dir.name}, indent=2) + "\n" + ) + return manifest + + +def _write_episode( + path: Path, + rows: list[dict[str, Any]], + *, + action_dim: int, + state_dim: int, + include_state: bool, +) -> None: + """Write one LeRobot episode parquet file. + + Args: + path: Destination episode parquet path. + rows: Converted LeRobot frame rows. + action_dim: Expected action vector dimension. + state_dim: Expected state vector dimension. + include_state: Whether rows carry real state vectors. + """ + path.parent.mkdir(parents=True, exist_ok=True) + states = ( + [row["state"] for row in rows] + if include_state + else [[0.0] * state_dim for _ in rows] + ) + columns = { + "observation.image": pa.array( + [{"bytes": row["image_bytes"], "path": None} for row in rows], + type=pa.struct([("bytes", pa.binary()), ("path", pa.string())]), + ), + "observation.state": pa.array( + states, + type=pa.list_(pa.float32(), state_dim), + ), + "action": pa.array( + [row["action"] for row in rows], + type=pa.list_(pa.float32(), action_dim), + ), + "timestamp": pa.array([row["timestamp"] for row in rows], type=pa.float64()), + "episode_index": pa.array( + [row["episode_index"] for row in rows], type=pa.int64() + ), + "frame_index": pa.array([row["frame_index"] for row in rows], type=pa.int64()), + "task_index": pa.array([row["task_index"] for row in rows], type=pa.int64()), + "latency": pa.array( + [row["latency_raw_frames"] for row in rows], type=pa.int64() + ), + "latency_raw_frames": pa.array( + [row["latency_raw_frames"] for row in rows], type=pa.int64() + ), + "done": pa.array([row["done"] for row in rows], type=pa.bool_()), + "reward": pa.array([row["reward"] for row in rows], type=pa.float32()), + # Per-row loss weight for KV-memory scheme-B supervision (clean_v1 event + # balancing without deleting rows). Uniform 1.0 when density is disabled. + "density_weight": pa.array( + [row["density_weight"] for row in rows], type=pa.float32() + ), + } + if rows and "issued_command_chunk" in rows[0]: + columns["issued_command_chunk"] = pa.array( + [row["issued_command_chunk"] for row in rows], + type=pa.list_(pa.list_(pa.float32(), action_dim), len(rows[0]["issued_command_chunk"])), + ) + for mask_key in ("action_prefix_mask", "action_loss_mask"): + if rows and mask_key in rows[0]: + columns[mask_key] = pa.array( + [row[mask_key] for row in rows], + type=pa.list_(pa.bool_(), len(rows[0][mask_key])), + ) + if rows and "action_id" in rows[0]: + columns["action_id"] = pa.array( + [row["action_id"] for row in rows], type=pa.int64() + ) + if rows and "context_image_bytes" in rows[0]: + image_type = pa.struct([("bytes", pa.binary()), ("path", pa.string())]) + columns["observation.context_images"] = pa.array( + [ + [{"bytes": value, "path": None} for value in row["context_image_bytes"]] + for row in rows + ], + type=pa.list_(image_type), + ) + table = pa.table(columns) + pq.write_table(table, path) + + +def _row_action_vector( + row: dict[str, Any], + *, + active_action_dim: int, + action_dim: int, +) -> list[float]: + values = [float(value) for value in row["action"]] + return [*values, *([0.0] * (action_dim - active_action_dim))] + + +def _write_metadata( + dataset_dir: Path, + *, + fps: float, + action_labels: list[str], + episode_lengths: list[int], + task_prompts: list[str], + image_shape: tuple[int, int, int], + state_dim: int, + state_labels: list[str], + context_image_count: int | None, +) -> None: + """Write minimal LeRobot metadata files for a split dataset. + + Args: + dataset_dir: LeRobot dataset split directory. + fps: Frame rate stored in metadata. + action_labels: Canonical action labels in vector order. + episode_lengths: Number of frames per converted episode. + task_prompts: Unique prompts used as LeRobot tasks. + """ + meta_dir = dataset_dir / "meta" + meta_dir.mkdir(parents=True, exist_ok=True) + action_dim = len(action_labels) + height, width, channels = image_shape + features = { + "observation.image": { + "dtype": "image", + "shape": [height, width, channels], + "names": ["height", "width", "channel"], + "video_info": {"video.fps": fps}, + }, + "observation.state": { + "dtype": "float32", + "shape": [state_dim], + "names": state_labels, + }, + "action": {"dtype": "float32", "shape": [action_dim], "names": action_labels}, + "timestamp": {"dtype": "float64", "shape": [1]}, + "episode_index": {"dtype": "int64", "shape": [1]}, + "frame_index": {"dtype": "int64", "shape": [1]}, + "task_index": {"dtype": "int64", "shape": [1]}, + "latency": {"dtype": "int64", "shape": [1]}, + } + if context_image_count is not None: + features["observation.context_images"] = { + "dtype": "image_sequence", + "shape": [context_image_count, height, width, channels], + "names": ["time", "height", "width", "channel"], + "video_info": {"video.fps": fps}, + } + (meta_dir / "info.json").write_text( + json.dumps( + { + "codebase_version": "v2.0", + "fps": fps, + "total_episodes": len(episode_lengths), + "total_frames": int(sum(episode_lengths)), + "total_tasks": len(task_prompts), + "total_videos": 0, + "total_chunks": (len(episode_lengths) + LE_ROBOT_CHUNK_SIZE - 1) + // LE_ROBOT_CHUNK_SIZE, + "chunks_size": LE_ROBOT_CHUNK_SIZE, + "data_path": "data/chunk-{episode_chunk:03d}/episode_{episode_index:06d}.parquet", + "video_path": ( + "videos/chunk-{episode_chunk:03d}/{video_key}/" + "episode_{episode_index:06d}.mp4" + ), + "features": features, + }, + indent=2, + ), + encoding="utf-8", + ) + (meta_dir / "modality.json").write_text( + json.dumps( + { + "state": { + "game_state": { + "start": 0, + "end": state_dim, + "dtype": "float32", + "absolute": True, + "original_key": "observation.state", + } + }, + "action": { + "button": { + "start": 0, + "end": action_dim, + "dtype": "float32", + "absolute": True, + "original_key": "action", + } + }, + "video": {"image": {"original_key": "observation.image"}}, + "annotation": { + "human.task_description": {"original_key": "task_index"} + }, + }, + indent=2, + ), + encoding="utf-8", + ) + write_jsonl( + meta_dir / "episodes.jsonl", + [ + {"episode_index": idx, "length": int(length)} + for idx, length in enumerate(episode_lengths) + ], + ) + write_jsonl( + meta_dir / "tasks.jsonl", + [ + {"task_index": idx, "task": prompt} + for idx, prompt in enumerate(task_prompts) + ], + ) + + +def _convert_split( + rows: list[dict[str, Any]], + output_dir: Path, + *, + env_name: str, + integration_name: str, + action_labels: list[str], + active_action_dim: int, + state_dim: int, + state_labels: list[str], + context_image_count: int | None, + image_shape: tuple[int, int, int], + fps: float, + reward_field: str, + state_normalization: dict[str, Any] | None, + include_state: bool, + density_weighting: bool = False, +) -> dict[str, Any]: + """Convert one canonical rollout split into a LeRobot dataset directory. + + Args: + rows: Canonical rollout rows from one split. + output_dir: Destination LeRobot split directory. + env_name: Canonical benchmark environment name. + action_labels: Canonical action labels in vector order. + fps: Frame rate used for timestamps. + reward_field: Source canonical reward column. + density_weighting: When True, write per-row ``density_weight`` from clean_v1's + event balancing (train split). When False, weights are uniform 1.0 (val + split keeps an unweighted held-out loss). + + Returns: + Split manifest summary. + """ + output_dir.mkdir(parents=True, exist_ok=True) + prompt_to_task_index: dict[str, int] = {} + task_prompts: list[str] = [] + episode_lengths: list[int] = [] + for new_episode_idx, episode_rows in enumerate(_episode_groups(rows)): + density_weights = ( + density_weights_clean_v1(env_name, episode_rows) + if density_weighting + and integration_name != "gymnasium" + else [1.0] * len(episode_rows) + ) + out_rows = [] + for frame_idx, row in enumerate(episode_rows): + # Enforce the canonical rollout contract; downstream must not read the legacy ``latency`` alias. + row["latency_raw_frames"] + prompt = str(row["prompt"]) + if prompt not in prompt_to_task_index: + prompt_to_task_index[prompt] = len(task_prompts) + task_prompts.append(prompt) + out_row = { + "image_bytes": _png_bytes(row["image"]), + "action": _row_action_vector( + row, + active_action_dim=active_action_dim, + action_dim=len(action_labels), + ), + "timestamp": frame_idx / fps, + "episode_index": new_episode_idx, + "frame_index": frame_idx, + "task_index": prompt_to_task_index[prompt], + "latency_raw_frames": int(row["latency_raw_frames"]), + # ``done`` is generated only for the LeRobot target format, not for canonical rollout rows. + "done": frame_idx == len(episode_rows) - 1, + "reward": float(row[reward_field]), + "density_weight": float(density_weights[frame_idx]), + } + if include_state: + state = np.asarray(row["state"], dtype=np.float32) + if state_normalization is not None: + state_min = np.asarray(state_normalization["min"], dtype=np.float32) + state_max = np.asarray(state_normalization["max"], dtype=np.float32) + state = min_max_normalize_state(state, state_min, state_max) + out_row["state"] = state.tolist() + if "issued_command_chunk" in row: + out_row["issued_command_chunk"] = row["issued_command_chunk"] + for mask_key in ("action_prefix_mask", "action_loss_mask"): + if mask_key in row: + out_row[mask_key] = row[mask_key] + if "action_id" in row: + out_row["action_id"] = int(row["action_id"]) + if "context_images" in row and row["context_images"] is not None: + out_row["context_image_bytes"] = [ + _png_bytes(image) for image in row["context_images"] + ] + out_rows.append(out_row) + episode_lengths.append(len(out_rows)) + episode_chunk = new_episode_idx // LE_ROBOT_CHUNK_SIZE + _write_episode( + output_dir + / f"data/chunk-{episode_chunk:03d}/episode_{new_episode_idx:06d}.parquet", + out_rows, + action_dim=len(action_labels), + state_dim=state_dim, + include_state=include_state, + ) + _write_metadata( + output_dir, + fps=fps, + action_labels=action_labels, + episode_lengths=episode_lengths, + task_prompts=task_prompts, + image_shape=image_shape, + state_dim=state_dim, + state_labels=state_labels, + context_image_count=context_image_count, + ) + if rows and "issued_command_chunk" in rows[0]: + info_path = output_dir / "meta/info.json" + info = json.loads(info_path.read_text()) + info["features"]["issued_command_chunk"] = { + "dtype": "float32", "shape": [len(rows[0]["issued_command_chunk"]), len(action_labels)], + "names": action_labels, + } + for mask_key in ("action_prefix_mask", "action_loss_mask"): + if mask_key in rows[0]: + info["features"][mask_key] = { + "dtype": "bool", "shape": [len(rows[0][mask_key])], "names": None, + } + info_path.write_text(json.dumps(info, indent=2) + "\n") + return { + "dataset_name": output_dir.name, + "env_name": env_name, + "episodes": len(episode_lengths), + "frames": int(sum(episode_lengths)), + "task_prompts": task_prompts, + } + + +def convert_canonical_rollout_to_lerobot( + raw_export_dir: Path | str, + output_dir: Path | str, + *, + force: bool = False, + action_carrier: str = "native", + normalize_state: bool = True, + write_starvla_handoff: bool = True, +) -> dict[str, Any]: + """Convert a canonical rollout export into train/val LeRobot datasets. + + Args: + raw_export_dir: Directory with canonical rollout parquet files and metadata. + output_dir: Destination LeRobot train dataset directory. + force: Remove existing train/val output directories before writing. + normalize_state: Apply the existing min-max state transform when the + canonical export includes a state vector. When disabled, an + explicit ``state`` field in canonical rows is copied as-is. + write_starvla_handoff: Write the existing StarVLA prompt/mixture handoff + files and manifest fields. + + Returns: + Manifest describing the converted train dataset and validation companion. + """ + raw_export_dir = Path(raw_export_dir) + output_dir = Path(output_dir) + metadata = json.loads( + (raw_export_dir / "metadata.json").read_text(encoding="utf-8") + ) + env_name = str(metadata["env_name"]) + integration_name = ( + metadata["integration_name"] if "integration_name" in metadata else env_name + ) + if metadata.get("action_history_decisions") == 8 and env_name in _HISTORY_STATE_DIMS: + return _convert_history_game_export( + raw_export_dir, output_dir, metadata=metadata, force=force + ) + if integration_name != "gymnasium": + return _convert_legacy_export( + raw_export_dir, + output_dir, + metadata=metadata, + force=force, + ) + action_layout = metadata["action_layout"] + active_action_labels = list(metadata["action_labels"]) + if action_layout not in ( + ACTION_LAYOUT_GYMNASIUM_DISCRETE_V1, + ACTION_LAYOUT_GYMNASIUM_CONTINUOUS_V1, + ACTION_LAYOUT_G1_SONIC_ABSOLUTE_JOINT_TARGET_V1, + ): + raise ValueError(f"Unsupported Gymnasium action_layout={action_layout!r}") + gymnasium_task = metadata["gymnasium_task"] + if env_name != gymnasium_task["task_name"]: + raise ValueError( + "Canonical Gymnasium env_name does not match gymnasium_task.task_name" + ) + continuous = action_layout in ( + ACTION_LAYOUT_GYMNASIUM_CONTINUOUS_V1, + ACTION_LAYOUT_G1_SONIC_ABSOLUTE_JOINT_TARGET_V1, + ) + if continuous and action_carrier == "native": + gymnasium_task = { + **gymnasium_task, + "action_space": { + key: value + for key, value in gymnasium_task["action_space"].items() + if key not in {"openvla_carrier_dim", "openvla_padding"} + }, + } + task_action_labels = ( + gymnasium_task["action_space"]["labels"] + if continuous + else gymnasium_task["action_labels"] + ) + if active_action_labels != task_action_labels: + raise ValueError( + "Canonical Gymnasium action_labels do not match gymnasium_task action-space labels" + ) + action_labels, active_action_dim, state_dim, state_labels = _action_carrier_spec( + active_action_labels, + action_carrier, + ) + train_rows = _load_rows(raw_export_dir, "train") + val_rows = _load_rows(raw_export_dir, "val") + raw_train_rows = train_rows + raw_val_rows = val_rows + admission_filtered = ( + action_layout == ACTION_LAYOUT_G1_SONIC_ABSOLUTE_JOINT_TARGET_V1 + and any("command_admitted" in row for row in [*train_rows, *val_rows]) + ) + if admission_filtered: + train_rows = [row for row in train_rows if row["command_admitted"]] + val_rows = [row for row in val_rows if row["command_admitted"]] + uses_state = "state_space" in gymnasium_task + if not normalize_state: + uses_state = any("state" in row for row in [*train_rows, *val_rows]) + if uses_state: + if "state_space" in gymnasium_task: + state_labels = list(gymnasium_task["state_space"]["labels"]) + state_dim = len(state_labels) + else: + state_row = next(row for row in [*train_rows, *val_rows] if "state" in row) + state_dim = len(state_row["state"]) + state_labels = [f"state_{index}" for index in range(state_dim)] + reward_field = str(metadata["reward_field"]) + rows_unit = metadata["rows_unit"] + if rows_unit != "decision_step": + raise ValueError(f"rows_unit must be 'decision_step', got {rows_unit!r}") + fps = metadata["obs_fps"] + obs_stride = metadata["obs_stride_raw_frames"] + val_output_dir = output_dir.with_name(f"{output_dir.name}__val") + if output_dir.exists() and force: + shutil.rmtree(output_dir) + if val_output_dir.exists() and force: + shutil.rmtree(val_output_dir) + state_normalization = None + if uses_state and normalize_state: + train_states = np.asarray([row["state"] for row in train_rows], dtype=np.float32) + state_normalization = { + "type": "min_max", + "min": train_states.min(axis=0).tolist(), + "max": train_states.max(axis=0).tolist(), + } + first_row = (raw_train_rows or raw_val_rows)[0] + image_source = first_row["image"] + with Image.open(io.BytesIO(_png_bytes(image_source))) as image: + width, height = image.size + image_shape = (height, width, 3) + context_image_count = ( + len(first_row["context_images"]) + if "context_images" in first_row and first_row["context_images"] is not None + else None + ) + train_manifest = _convert_split( + train_rows, + output_dir, + env_name=env_name, + integration_name=integration_name, + action_labels=action_labels, + active_action_dim=active_action_dim, + state_dim=state_dim, + state_labels=state_labels, + context_image_count=context_image_count, + image_shape=image_shape, + fps=fps, + reward_field=reward_field, + state_normalization=state_normalization, + include_state=uses_state, + density_weighting=True, + ) + val_manifest = _convert_split( + val_rows, + val_output_dir, + env_name=env_name, + integration_name=integration_name, + action_labels=action_labels, + active_action_dim=active_action_dim, + state_dim=state_dim, + state_labels=state_labels, + context_image_count=context_image_count, + image_shape=image_shape, + fps=fps, + reward_field=reward_field, + state_normalization=state_normalization, + include_state=uses_state, + density_weighting=False, + ) + if "label_provenance" in metadata: + train_info = json.loads((output_dir / "meta/info.json").read_text()) + for directory in (output_dir, val_output_dir): + info_path = directory / "meta/info.json" + info = json.loads(info_path.read_text()) + info["label_provenance"] = metadata["label_provenance"] + for key in ("issued_command_chunk", "action_prefix_mask", "action_loss_mask"): + if key in train_info["features"]: + info["features"][key] = train_info["features"][key] + info_path.write_text(json.dumps(info, indent=2) + "\n") + handoff_contract = { + "format": ( + "lerobot_v2_image_parquet" + if continuous + else "starvla_lerobot_v2_image_parquet" + ), + "source": str(raw_export_dir), + "integration_name": integration_name, + "task_name": env_name, + "action_layout": str(metadata["action_layout"]), + "action_labels": active_action_labels, + "carrier_action_labels": action_labels, + "action_dim": len(action_labels), + "active_action_dim": active_action_dim, + "action_carrier": action_carrier, + "rows_unit": rows_unit, + "obs_fps": fps, + "obs_stride_raw_frames": obs_stride, + "uses_state": uses_state, + "state_dim": state_dim, + "state_labels": state_labels, + "state_normalization": state_normalization, + } + if admission_filtered: + handoff_contract.update( + { + "admission_filter": "command_admitted", + "raw_train_frames": len(raw_train_rows), + "raw_validation_frames": len(raw_val_rows), + } + ) + if write_starvla_handoff: + handoff_contract["robot_type"] = _robot_type_for_env(env_name, integration_name) + prompt_map_path, mixture_path = _write_starvla_handoff_files( + output_dir, + val_output_dir, + env_name=env_name, + integration_name=integration_name, + rows=[*train_rows, *val_rows], + ) + handoff_contract.update( + { + "latency_prompt_map_path": str(prompt_map_path), + "custom_mixtures_path": str(mixture_path), + } + ) + if gymnasium_task is not None: + handoff_contract["gymnasium_task"] = gymnasium_task + manifest = { + **train_manifest, + **handoff_contract, + "validation_dataset_name": val_output_dir.name, + "validation_episodes": val_manifest["episodes"], + "validation_frames": val_manifest["frames"], + } + validation_manifest = { + **val_manifest, + **handoff_contract, + "training_dataset_name": output_dir.name, + } + (output_dir / "manifest.json").write_text( + json.dumps(manifest, indent=2), encoding="utf-8" + ) + (val_output_dir / "manifest.json").write_text( + json.dumps(validation_manifest, indent=2), encoding="utf-8" + ) + return manifest diff --git a/latency-aware/deadly-corridor/mean_rtx3090_s42_8500ep_v1/files.sha256.json b/latency-aware/deadly-corridor/mean_rtx3090_s42_8500ep_v1/files.sha256.json new file mode 100644 index 0000000000000000000000000000000000000000..95ea593a0ac0d30535e49a83cd67120eecdc6c5f --- /dev/null +++ b/latency-aware/deadly-corridor/mean_rtx3090_s42_8500ep_v1/files.sha256.json @@ -0,0 +1,106 @@ +{ + "README.md": { + "bytes": 812, + "sha256": "6262b308d99d95e8c2bd276ca31320d86797b0ffd116fc2a3b201e806f5db770" + }, + "demonstrations/conversion/README.md": { + "bytes": 412, + "sha256": "b0cc5a2a0681575c7fd23150d5b3211f65236832b4f9c715bd0b84404d35ccdf" + }, + "demonstrations/conversion/convert_to_lerobot.py": { + "bytes": 230, + "sha256": "4570ac62854befced86409de1c911b857d6a852202a3679fdfe94555ad84b3e7" + }, + "demonstrations/conversion/expected/_generated_mixtures/lerobot.json": { + "bytes": 198, + "sha256": "63130f38b8f5b844b3b5be0ee8aaff40c20e3aa9c48e819d7a197385f69b02be" + }, + "demonstrations/conversion/expected/lerobot/manifest.json": { + "bytes": 7393, + "sha256": "7168be7a6a3829b581c15a81af4afb5509bab7c466e05b8a62cfe24c80c72bd9" + }, + "demonstrations/conversion/expected/lerobot/meta/episodes.jsonl": { + "bytes": 289543, + "sha256": "ff0fd2a40913dda90a740058377fd37804d43e2f4ef8563538fc4a279db1c0bc" + }, + "demonstrations/conversion/expected/lerobot/meta/info.json": { + "bytes": 5186, + "sha256": "d6835c3906d92967a241c6efaaa3567f81403361a9813dfcdf943e74f78da19b" + }, + "demonstrations/conversion/expected/lerobot/meta/modality.json": { + "bytes": 515, + "sha256": "ea051012ef748ce22610be3a8841f92fec2b58c01ed4f0bab5f4ce511a1ac0d5" + }, + "demonstrations/conversion/expected/lerobot/meta/tasks.jsonl": { + "bytes": 352, + "sha256": "dd338eb777e99789317e65f69178a6bb6dc29e666910f9aae573a63cebbb78b7" + }, + "demonstrations/conversion/expected/lerobot__val/manifest.json": { + "bytes": 7329, + "sha256": "f1c5fa37f1dc59c3b514864678fe75fc6f407e668092d8f9397d6b32586c8275" + }, + "demonstrations/conversion/expected/lerobot__val/meta/episodes.jsonl": { + "bytes": 31333, + "sha256": "9b73fe52d8da21a5baf25628f56efaa23f5fcf4cd123be19d2aec4b399c1783a" + }, + "demonstrations/conversion/expected/lerobot__val/meta/info.json": { + "bytes": 5184, + "sha256": "4cc3b2dd637b41743311b56e862be6f8641b2deeaae9781da50e05094ca2e353" + }, + "demonstrations/conversion/expected/lerobot__val/meta/modality.json": { + "bytes": 515, + "sha256": "ea051012ef748ce22610be3a8841f92fec2b58c01ed4f0bab5f4ce511a1ac0d5" + }, + "demonstrations/conversion/expected/lerobot__val/meta/tasks.jsonl": { + "bytes": 352, + "sha256": "dd338eb777e99789317e65f69178a6bb6dc29e666910f9aae573a63cebbb78b7" + }, + "demonstrations/raw/filter_report.json": { + "bytes": 335, + "sha256": "30d84915ce8b56cc3012c3bc8520af57cf241d770bf8b43496a5e30373500ff0" + }, + "demonstrations/raw/metadata.json": { + "bytes": 15375, + "sha256": "29ab118d9439fdc55454619965e0cbfc774bb34043570821243a0a74d9f25bf9" + }, + "demonstrations/raw/train.parquet": { + "bytes": 4950850571, + "sha256": "3e3473cf7eeece0b3e107f516869b136159fd197ff4a8667a47e128cf336b9e2" + }, + "demonstrations/raw/val.parquet": { + "bytes": 543289255, + "sha256": "68456699331b2362acf90f859b9d5b379e844a6d38d6047c9f67d2cceb79b269" + }, + "demonstrations/validation/collection.complete.json": { + "bytes": 452, + "sha256": "244d67ae4831598951770ecf4b95f8d89799305adfa7eb743095541e58277a62" + }, + "demonstrations/validation/dataset.audit.json": { + "bytes": 784, + "sha256": "57f59e18bf29b9c183ff3572f33f83b49344648f859925557df55f7b6c920172" + }, + "evidence/collect_sf.py": { + "bytes": 2650, + "sha256": "883cdfc33382e3cdbb628124a4eb7c736ecaa6bdb3757d9fca00a21f4586d037" + }, + "evidence/collection-command.json": { + "bytes": 688, + "sha256": "93559e46898621f9336c2f468c1654d253cd14c540fb934e3063f30dcae6df14" + }, + "evidence/doom-mean-reset.patch": { + "bytes": 9569, + "sha256": "4113989e16c547bd1f02f5f10f285c4d511be3a3c18d10a0aa3316bd8ff3175d" + }, + "evidence/profile.source.json": { + "bytes": 2115, + "sha256": "bbfb93cef9f63750a9a159ec10a93a9fc6c25e4cb0cac3b39532663b4dc11eba" + }, + "evidence/starvla_lerobot.source.py": { + "bytes": 42762, + "sha256": "c3aae215af7c7f04051d7a2731a2007727f7ac3426fbc8d2aad2f6a276ae842b" + }, + "provenance.json": { + "bytes": 3762, + "sha256": "f38b40bafc78d3bfacb4bf131fa18b7a009961dd20a38f84ccaa5aef2bef7926" + } +} diff --git a/latency-aware/deadly-corridor/mean_rtx3090_s42_8500ep_v1/provenance.json b/latency-aware/deadly-corridor/mean_rtx3090_s42_8500ep_v1/provenance.json new file mode 100644 index 0000000000000000000000000000000000000000..5825eb4ec5d4e4d7a0a08c04d701b2ec2acfd897 --- /dev/null +++ b/latency-aware/deadly-corridor/mean_rtx3090_s42_8500ep_v1/provenance.json @@ -0,0 +1,101 @@ +{ + "task": "deadly_corridor", + "downstream_model": "QwenOFT", + "path_family_alias": "QwenOFT", + "teacher": { + "repo_id": "latency-sensitive-bench/benchmark-models", + "revision": "3ad793d76cee71c9043e696f7042d89c7cf7f7be", + "path": "latency-aware/deadly-corridor/small-policy/sample-factory-qwenoft-rtx3090-mean-hist8-capacity1-native35-s42-v1", + "local_checkpoint": "/home/ubuntu/lzj/mean-profiling/deadly_corridor/teacher-publication/checkpoint.pth", + "status": "complete_verified", + "completed_env_steps": 25006080, + "selected_env_steps": 22239232, + "selected_source_checkpoint": "/home/ubuntu/lzj/mean-profiling/deadly_corridor/teacher/deadly_corridor_mean_3090_s42/checkpoint_p0/best_000043436_22239232_reward_2141.427.pth", + "tensor_count": 25, + "weights_only_export": true, + "tensor_equality_verified": true, + "hash_readback_verified": true, + "url": "https://huggingface.co/latency-sensitive-bench/benchmark-models/tree/3ad793d76cee71c9043e696f7042d89c7cf7f7be/latency-aware/deadly-corridor/small-policy/sample-factory-qwenoft-rtx3090-mean-hist8-capacity1-native35-s42-v1", + "completed_utc": "2026-09-30T14:07:34.452490+00:00" + }, + "data": { + "task": "deadly_corridor", + "teacher_checkpoint": "/home/ubuntu/lzj/mean-profiling/deadly_corridor/teacher-publication/checkpoint.pth", + "teacher_sha256": "e631e9017634f799860c581761b5fa48601109393c11d20a35187c6f75d14b88", + "episodes": 8500, + "train_episodes": 7650, + "val_episodes": 850, + "train_rows": 313925, + "val_rows": 34465, + "fixed_latency_ms": 73.69250777493353, + "action_horizon": 1, + "unique_seeds": 8500, + "state_dim": 120 + }, + "runtime": { + "repo": "c3c6a39365a151e9b7a5e215452fd64e957c2b29", + "starvla": "ccca13c5177fe3d3c884b6e2de4965d916016649", + "runtime_fixes": [ + "mean-profile-preparation.patch", + "gym-language-contract.patch", + "loader-spawn-cache.patch", + "loader-spawn-test.patch", + "doom-mean-reset.patch" + ], + "pytorch3d": { + "revision": "33824be3cbc87a7dd1db0f6a9a9de9ac81b2d0ba", + "build": "transforms-only, no native render extension; QwenOFT uses transforms only" + }, + "decord": { + "version": "0.6.0", + "build": "official source CPU decoder CP310", + "wheel_sha256": "e193b356b1e984b4eff08d23b62e482c2c9e5037a6efdc0b1af47079ae2e4c47" + } + }, + "profile": { + "mean_ms": 73.69250777493353, + "profile": "/home/ubuntu/lzj/profiles/published/profiles/openvla/1x-rtx3090/deadly_corridor/instance_a5037b165aa0cedc/profile.json", + "sha256": "bbfb93cef9f63750a9a159ec10a93a9fc6c25e4cb0cac3b39532663b4dc11eba" + }, + "audit": { + "task": "deadly_corridor", + "episodes": 8500, + "train_episodes": 7650, + "val_episodes": 850, + "unique_seeds": 8500, + "seed_range": [ + 42, + 8541 + ], + "split_intersection": 0, + "train_rows": 313925, + "val_rows": 34465, + "rows": 348390, + "state_dim": 120, + "history_decisions": 8, + "state_binary": true, + "action_dim": 7, + "source_joint_actions": 54, + "converted_native_binary_actions": true, + "finite_numeric": true, + "image_shape": [ + 120, + 160, + 3 + ], + "decoded_image_samples": 6, + "mean_ms": 73.69250777493353, + "effective_latency_raw_frames": 3, + "effective_latency_ms": 85.71428571428571, + "env_fps": 35, + "obs_fps": 8.75, + "raw_max_steps": 3600, + "raw_parquet_bytes": 5494139826, + "converted_all_episode_row_counts_verified": true + }, + "condition": "fixed_mean_from_RTX3090_profile", + "raw_format": "canonical embeddedPNG parquet schema2", + "training_data_format": "converted LeRobot v2", + "vla_training_status": "not started", + "closed_loop_evaluation_status": "pending" +} diff --git a/latency-aware/flappy/mean_rtx3090_s42_200ep_v1/README.md b/latency-aware/flappy/mean_rtx3090_s42_200ep_v1/README.md new file mode 100644 index 0000000000000000000000000000000000000000..788eb8d4031d955cd83428b544a68456dacbcbec --- /dev/null +++ b/latency-aware/flappy/mean_rtx3090_s42_200ep_v1/README.md @@ -0,0 +1,15 @@ +# Flappy fixed-mean demonstrations + +This independent mean experiment uses fixed **75.87417450998383 ms** latency, quantized to **1 raw frame / 100 ms at 10 Hz**. `profile_latency` is the archive condition namespace; data use a fixed mean, not temporal replay or distribution sampling. The mean came from an OpenVLA RTX3090 measured profile; the intended downstream VLA is QwenOFT H1. + +200 unique-seed episodes (42..241), 18,000 decision rows each: train180 / 3,240,000 rows, validation20 / 360,000 rows. Deterministic teacher collection, split_seed0, validation fraction0.1, no filtering, raw RGB512x288, eight-decision binary24D causal state, exact2D one-hot actions. Both canonical Parquet files embed the PNG bytes. + +- [Raw demonstrations](demonstrations/raw/) +- [CPU audit](demonstrations/validation/dataset.audit.json) and [collection completion](demonstrations/validation/collection.complete.json) +- [Training conversion recipe](demonstrations/conversion/README.md) +- [Frozen source and teacher/profile provenance](evidence/provenance/source.json) +- [Publication checksum manifest](evidence/files.sha256.json) + +Raw data and LeRobot conversion completed; the CPU audit passed. This records data validity, not closed-loop model performance. The historical collection wrapper exit1 was a post-export read of the cleaned temporary probe_state; finalize-only exit0 repaired it without recollection or changing these canonical Parquet files. Full numeric raw validation covered all rows; raw-to-converted pixels/row identity was checked on30 sampled frames. See the audit for scope and limitations. + +Only canonical raw demonstrations are distributed, following this repository's policy. LeRobot image payloads remain on the training server and can be regenerated using the frozen conversion contract. Model weights live in `latency-sensitive-bench/benchmark-models` at the pinned teacher revision in provenance. No old temporal artifact is overwritten. diff --git a/latency-aware/flappy/mean_rtx3090_s42_200ep_v1/demonstrations/conversion/README.md b/latency-aware/flappy/mean_rtx3090_s42_200ep_v1/demonstrations/conversion/README.md new file mode 100644 index 0000000000000000000000000000000000000000..f8b6f62fdbc0bedac7a005746eabd4350b036a59 --- /dev/null +++ b/latency-aware/flappy/mean_rtx3090_s42_200ep_v1/demonstrations/conversion/README.md @@ -0,0 +1,13 @@ +# Reproduce the training data + +Use latency-sensitive-bench at `c3c6a39365a151e9b7a5e215452fd64e957c2b29`, StarVLA at `ccca13c5177fe3d3c884b6e2de4965d916016649`, and the exact Python/package versions in `../../evidence/provenance/source.json`. The frozen converter and registry source snapshots are alongside that manifest. The collection streaming patch only changes bounded image writing; apply it when reproducing collection, not conversion. + +From the checked-out latency-sensitive-bench source, expose its package on PYTHONPATH and run: + +```bash +PYTHONPATH=/path/to/latency-sensitive-bench python convert_to_lerobot.py --raw /path/to/demonstrations/raw --output /path/to/data/lerobot +``` + +This creates `lerobot` (180 episodes / 3,240,000 rows), companion `lerobot__val` (20 episodes / 360,000 rows), and `_generated_mixtures/lerobot.json`. Native actions are exact 2D one-hot; binary eight-decision histories are 24D. The `expected/` metadata records the completed local conversion. Its absolute paths document the original host and must be regenerated for a new output location. Raw PNG image bytes are embedded in the canonical Parquet; separate images are not needed. + +The reference training config is `../../evidence/provenance/vla.training-config.reference.yaml`: QwenOFT H1, action_dim=2, state_dim=24, raw prompt, native action carrier, 224x224 model resize, global batch128 (micro16 x accumulation8), seed42, 5,000 configured updates. It is a configuration reference, not proof of completed VLA training. Point dataset paths and the base-model path to the new host after conversion. The archived registry snapshot preserves the exact Flappy modality contract. diff --git a/latency-aware/flappy/mean_rtx3090_s42_200ep_v1/demonstrations/conversion/convert_to_lerobot.py b/latency-aware/flappy/mean_rtx3090_s42_200ep_v1/demonstrations/conversion/convert_to_lerobot.py new file mode 100644 index 0000000000000000000000000000000000000000..f0ed54d13fcef3f55732a5fb53a26328f2fc27d4 --- /dev/null +++ b/latency-aware/flappy/mean_rtx3090_s42_200ep_v1/demonstrations/conversion/convert_to_lerobot.py @@ -0,0 +1,9 @@ +from pathlib import Path +import argparse +from latency_bench.data.exporters.starvla_lerobot import convert_canonical_rollout_to_lerobot + +parser = argparse.ArgumentParser() +parser.add_argument("--raw", type=Path, required=True) +parser.add_argument("--output", type=Path, required=True) +args = parser.parse_args() +convert_canonical_rollout_to_lerobot(args.raw, args.output, action_carrier="native") diff --git a/latency-aware/flappy/mean_rtx3090_s42_200ep_v1/demonstrations/conversion/expected/_generated_mixtures/lerobot.json b/latency-aware/flappy/mean_rtx3090_s42_200ep_v1/demonstrations/conversion/expected/_generated_mixtures/lerobot.json new file mode 100644 index 0000000000000000000000000000000000000000..90240ec6f4201c9e5379b57ecae93e73500796a1 --- /dev/null +++ b/latency-aware/flappy/mean_rtx3090_s42_200ep_v1/demonstrations/conversion/expected/_generated_mixtures/lerobot.json @@ -0,0 +1,16 @@ +{ + "lerobot": [ + [ + "lerobot", + 1.0, + "rl_games_flappy" + ] + ], + "lerobot__val": [ + [ + "lerobot__val", + 1.0, + "rl_games_flappy" + ] + ] +} \ No newline at end of file diff --git a/latency-aware/flappy/mean_rtx3090_s42_200ep_v1/demonstrations/conversion/expected/lerobot/latency_prompt_map.json b/latency-aware/flappy/mean_rtx3090_s42_200ep_v1/demonstrations/conversion/expected/lerobot/latency_prompt_map.json new file mode 100644 index 0000000000000000000000000000000000000000..465788c1cf15573a239a3146b6aca2bb8b5deaa4 --- /dev/null +++ b/latency-aware/flappy/mean_rtx3090_s42_200ep_v1/demonstrations/conversion/expected/lerobot/latency_prompt_map.json @@ -0,0 +1,7 @@ +{ + "1": { + "prompt": "You are playing Flappy Bird. Pass through the pipe gaps and stay alive. Choose the action: NOOP, FLAP.", + "latency_raw_frames": 1, + "latency_ms": 100.0 + } +} \ No newline at end of file diff --git a/latency-aware/flappy/mean_rtx3090_s42_200ep_v1/demonstrations/conversion/expected/lerobot/manifest.json b/latency-aware/flappy/mean_rtx3090_s42_200ep_v1/demonstrations/conversion/expected/lerobot/manifest.json new file mode 100644 index 0000000000000000000000000000000000000000..34f74342c9d2c900c37313a53a88d65e05e6125f --- /dev/null +++ b/latency-aware/flappy/mean_rtx3090_s42_200ep_v1/demonstrations/conversion/expected/lerobot/manifest.json @@ -0,0 +1,111 @@ +{ + "dataset_name": "lerobot", + "env_name": "flappy", + "episodes": 180, + "frames": 3240000, + "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": "/data/ubuntu/lzj/mean-profiling/flappy/data/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": "/data/ubuntu/lzj/mean-profiling/flappy/data/lerobot/latency_prompt_map.json", + "custom_mixtures_path": "/data/ubuntu/lzj/mean-profiling/flappy/data/_generated_mixtures/lerobot.json", + "validation_dataset_name": "lerobot__val", + "validation_episodes": 20, + "validation_frames": 360000 +} diff --git a/latency-aware/flappy/mean_rtx3090_s42_200ep_v1/demonstrations/conversion/expected/lerobot/meta/episodes.jsonl b/latency-aware/flappy/mean_rtx3090_s42_200ep_v1/demonstrations/conversion/expected/lerobot/meta/episodes.jsonl new file mode 100644 index 0000000000000000000000000000000000000000..d4cfdea0f4cb2c32438c8907c01c9226eab75699 --- /dev/null +++ b/latency-aware/flappy/mean_rtx3090_s42_200ep_v1/demonstrations/conversion/expected/lerobot/meta/episodes.jsonl @@ -0,0 +1,180 @@ +{"episode_index": 0, "length": 18000} +{"episode_index": 1, "length": 18000} +{"episode_index": 2, "length": 18000} +{"episode_index": 3, "length": 18000} +{"episode_index": 4, "length": 18000} +{"episode_index": 5, "length": 18000} +{"episode_index": 6, "length": 18000} +{"episode_index": 7, "length": 18000} +{"episode_index": 8, "length": 18000} +{"episode_index": 9, "length": 18000} +{"episode_index": 10, "length": 18000} +{"episode_index": 11, "length": 18000} +{"episode_index": 12, "length": 18000} +{"episode_index": 13, "length": 18000} +{"episode_index": 14, "length": 18000} +{"episode_index": 15, "length": 18000} +{"episode_index": 16, "length": 18000} +{"episode_index": 17, "length": 18000} +{"episode_index": 18, "length": 18000} +{"episode_index": 19, "length": 18000} +{"episode_index": 20, "length": 18000} +{"episode_index": 21, "length": 18000} +{"episode_index": 22, "length": 18000} +{"episode_index": 23, "length": 18000} +{"episode_index": 24, "length": 18000} +{"episode_index": 25, "length": 18000} +{"episode_index": 26, "length": 18000} +{"episode_index": 27, "length": 18000} +{"episode_index": 28, "length": 18000} +{"episode_index": 29, "length": 18000} +{"episode_index": 30, "length": 18000} +{"episode_index": 31, "length": 18000} +{"episode_index": 32, "length": 18000} +{"episode_index": 33, "length": 18000} +{"episode_index": 34, "length": 18000} +{"episode_index": 35, "length": 18000} +{"episode_index": 36, "length": 18000} +{"episode_index": 37, "length": 18000} +{"episode_index": 38, "length": 18000} +{"episode_index": 39, "length": 18000} +{"episode_index": 40, "length": 18000} +{"episode_index": 41, "length": 18000} +{"episode_index": 42, "length": 18000} +{"episode_index": 43, "length": 18000} +{"episode_index": 44, "length": 18000} +{"episode_index": 45, "length": 18000} +{"episode_index": 46, "length": 18000} +{"episode_index": 47, "length": 18000} +{"episode_index": 48, "length": 18000} +{"episode_index": 49, "length": 18000} +{"episode_index": 50, "length": 18000} +{"episode_index": 51, "length": 18000} +{"episode_index": 52, "length": 18000} +{"episode_index": 53, "length": 18000} +{"episode_index": 54, "length": 18000} +{"episode_index": 55, "length": 18000} +{"episode_index": 56, "length": 18000} +{"episode_index": 57, "length": 18000} +{"episode_index": 58, "length": 18000} +{"episode_index": 59, "length": 18000} +{"episode_index": 60, "length": 18000} +{"episode_index": 61, "length": 18000} +{"episode_index": 62, "length": 18000} +{"episode_index": 63, "length": 18000} +{"episode_index": 64, "length": 18000} +{"episode_index": 65, "length": 18000} +{"episode_index": 66, "length": 18000} +{"episode_index": 67, "length": 18000} +{"episode_index": 68, "length": 18000} +{"episode_index": 69, "length": 18000} +{"episode_index": 70, "length": 18000} +{"episode_index": 71, "length": 18000} +{"episode_index": 72, "length": 18000} +{"episode_index": 73, "length": 18000} +{"episode_index": 74, "length": 18000} +{"episode_index": 75, "length": 18000} +{"episode_index": 76, "length": 18000} +{"episode_index": 77, "length": 18000} +{"episode_index": 78, "length": 18000} +{"episode_index": 79, "length": 18000} +{"episode_index": 80, "length": 18000} +{"episode_index": 81, "length": 18000} +{"episode_index": 82, "length": 18000} +{"episode_index": 83, "length": 18000} +{"episode_index": 84, "length": 18000} +{"episode_index": 85, "length": 18000} +{"episode_index": 86, "length": 18000} +{"episode_index": 87, "length": 18000} +{"episode_index": 88, "length": 18000} +{"episode_index": 89, "length": 18000} +{"episode_index": 90, "length": 18000} +{"episode_index": 91, "length": 18000} +{"episode_index": 92, "length": 18000} +{"episode_index": 93, "length": 18000} +{"episode_index": 94, "length": 18000} +{"episode_index": 95, "length": 18000} +{"episode_index": 96, "length": 18000} +{"episode_index": 97, "length": 18000} +{"episode_index": 98, "length": 18000} +{"episode_index": 99, "length": 18000} +{"episode_index": 100, "length": 18000} +{"episode_index": 101, "length": 18000} +{"episode_index": 102, "length": 18000} +{"episode_index": 103, "length": 18000} +{"episode_index": 104, "length": 18000} +{"episode_index": 105, "length": 18000} +{"episode_index": 106, "length": 18000} +{"episode_index": 107, "length": 18000} +{"episode_index": 108, "length": 18000} +{"episode_index": 109, "length": 18000} +{"episode_index": 110, "length": 18000} +{"episode_index": 111, "length": 18000} +{"episode_index": 112, "length": 18000} +{"episode_index": 113, "length": 18000} +{"episode_index": 114, "length": 18000} +{"episode_index": 115, "length": 18000} +{"episode_index": 116, "length": 18000} +{"episode_index": 117, "length": 18000} +{"episode_index": 118, "length": 18000} +{"episode_index": 119, "length": 18000} +{"episode_index": 120, "length": 18000} +{"episode_index": 121, "length": 18000} +{"episode_index": 122, "length": 18000} +{"episode_index": 123, "length": 18000} +{"episode_index": 124, "length": 18000} +{"episode_index": 125, "length": 18000} +{"episode_index": 126, "length": 18000} +{"episode_index": 127, "length": 18000} +{"episode_index": 128, "length": 18000} +{"episode_index": 129, "length": 18000} +{"episode_index": 130, "length": 18000} +{"episode_index": 131, "length": 18000} +{"episode_index": 132, "length": 18000} +{"episode_index": 133, "length": 18000} +{"episode_index": 134, "length": 18000} +{"episode_index": 135, "length": 18000} +{"episode_index": 136, "length": 18000} +{"episode_index": 137, "length": 18000} +{"episode_index": 138, "length": 18000} +{"episode_index": 139, "length": 18000} +{"episode_index": 140, "length": 18000} +{"episode_index": 141, "length": 18000} +{"episode_index": 142, "length": 18000} +{"episode_index": 143, "length": 18000} +{"episode_index": 144, "length": 18000} +{"episode_index": 145, "length": 18000} +{"episode_index": 146, "length": 18000} +{"episode_index": 147, "length": 18000} +{"episode_index": 148, "length": 18000} +{"episode_index": 149, "length": 18000} +{"episode_index": 150, "length": 18000} +{"episode_index": 151, "length": 18000} +{"episode_index": 152, "length": 18000} +{"episode_index": 153, "length": 18000} +{"episode_index": 154, "length": 18000} +{"episode_index": 155, "length": 18000} +{"episode_index": 156, "length": 18000} +{"episode_index": 157, "length": 18000} +{"episode_index": 158, "length": 18000} +{"episode_index": 159, "length": 18000} +{"episode_index": 160, "length": 18000} +{"episode_index": 161, "length": 18000} +{"episode_index": 162, "length": 18000} +{"episode_index": 163, "length": 18000} +{"episode_index": 164, "length": 18000} +{"episode_index": 165, "length": 18000} +{"episode_index": 166, "length": 18000} +{"episode_index": 167, "length": 18000} +{"episode_index": 168, "length": 18000} +{"episode_index": 169, "length": 18000} +{"episode_index": 170, "length": 18000} +{"episode_index": 171, "length": 18000} +{"episode_index": 172, "length": 18000} +{"episode_index": 173, "length": 18000} +{"episode_index": 174, "length": 18000} +{"episode_index": 175, "length": 18000} +{"episode_index": 176, "length": 18000} +{"episode_index": 177, "length": 18000} +{"episode_index": 178, "length": 18000} +{"episode_index": 179, "length": 18000} diff --git a/latency-aware/flappy/mean_rtx3090_s42_200ep_v1/demonstrations/conversion/expected/lerobot/meta/info.json b/latency-aware/flappy/mean_rtx3090_s42_200ep_v1/demonstrations/conversion/expected/lerobot/meta/info.json new file mode 100644 index 0000000000000000000000000000000000000000..b0dd07442e6b4b29a7aec730386ed865e747375e --- /dev/null +++ b/latency-aware/flappy/mean_rtx3090_s42_200ep_v1/demonstrations/conversion/expected/lerobot/meta/info.json @@ -0,0 +1,102 @@ +{ + "codebase_version": "v2.0", + "fps": 10.0, + "total_episodes": 180, + "total_frames": 3240000, + "total_tasks": 1, + "total_videos": 0, + "total_chunks": 1, + "chunks_size": 1000, + "data_path": "data/chunk-{episode_chunk:03d}/episode_{episode_index:06d}.parquet", + "video_path": "videos/chunk-{episode_chunk:03d}/{video_key}/episode_{episode_index:06d}.mp4", + "features": { + "observation.image": { + "dtype": "image", + "shape": [ + 512, + 288, + 3 + ], + "names": [ + "height", + "width", + "channel" + ], + "video_info": { + "video.fps": 10.0 + } + }, + "observation.state": { + "dtype": "float32", + "shape": [ + 24 + ], + "names": [ + "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" + ] + }, + "action": { + "dtype": "float32", + "shape": [ + 2 + ], + "names": [ + "NOOP", + "FLAP" + ] + }, + "timestamp": { + "dtype": "float64", + "shape": [ + 1 + ] + }, + "episode_index": { + "dtype": "int64", + "shape": [ + 1 + ] + }, + "frame_index": { + "dtype": "int64", + "shape": [ + 1 + ] + }, + "task_index": { + "dtype": "int64", + "shape": [ + 1 + ] + }, + "latency": { + "dtype": "int64", + "shape": [ + 1 + ] + } + } +} \ No newline at end of file diff --git a/latency-aware/flappy/mean_rtx3090_s42_200ep_v1/demonstrations/conversion/expected/lerobot/meta/modality.json b/latency-aware/flappy/mean_rtx3090_s42_200ep_v1/demonstrations/conversion/expected/lerobot/meta/modality.json new file mode 100644 index 0000000000000000000000000000000000000000..9b611ad7557a5feabdc80e1c2f11a706a99bb45c --- /dev/null +++ b/latency-aware/flappy/mean_rtx3090_s42_200ep_v1/demonstrations/conversion/expected/lerobot/meta/modality.json @@ -0,0 +1,30 @@ +{ + "state": { + "game_state": { + "start": 0, + "end": 24, + "dtype": "float32", + "absolute": true, + "original_key": "observation.state" + } + }, + "action": { + "button": { + "start": 0, + "end": 2, + "dtype": "float32", + "absolute": true, + "original_key": "action" + } + }, + "video": { + "image": { + "original_key": "observation.image" + } + }, + "annotation": { + "human.action.task_description": { + "original_key": "task_index" + } + } +} diff --git a/latency-aware/flappy/mean_rtx3090_s42_200ep_v1/demonstrations/conversion/expected/lerobot/meta/tasks.jsonl b/latency-aware/flappy/mean_rtx3090_s42_200ep_v1/demonstrations/conversion/expected/lerobot/meta/tasks.jsonl new file mode 100644 index 0000000000000000000000000000000000000000..8a4acdb4011cc11914f414e2c5b7bf68740f8781 --- /dev/null +++ b/latency-aware/flappy/mean_rtx3090_s42_200ep_v1/demonstrations/conversion/expected/lerobot/meta/tasks.jsonl @@ -0,0 +1 @@ +{"task": "You are playing Flappy Bird. Pass through the pipe gaps and stay alive. Choose the action: NOOP, FLAP.", "task_index": 0} diff --git a/latency-aware/flappy/mean_rtx3090_s42_200ep_v1/demonstrations/conversion/expected/lerobot__val/manifest.json b/latency-aware/flappy/mean_rtx3090_s42_200ep_v1/demonstrations/conversion/expected/lerobot__val/manifest.json new file mode 100644 index 0000000000000000000000000000000000000000..2c8fd2d341f04397e8243fb451f8ad74dfbb3a0d --- /dev/null +++ b/latency-aware/flappy/mean_rtx3090_s42_200ep_v1/demonstrations/conversion/expected/lerobot__val/manifest.json @@ -0,0 +1,109 @@ +{ + "dataset_name": "lerobot__val", + "env_name": "flappy", + "episodes": 20, + "frames": 360000, + "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": "/data/ubuntu/lzj/mean-profiling/flappy/data/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": "/data/ubuntu/lzj/mean-profiling/flappy/data/lerobot/latency_prompt_map.json", + "custom_mixtures_path": "/data/ubuntu/lzj/mean-profiling/flappy/data/_generated_mixtures/lerobot.json", + "training_dataset_name": "lerobot" +} diff --git a/latency-aware/flappy/mean_rtx3090_s42_200ep_v1/demonstrations/conversion/expected/lerobot__val/meta/episodes.jsonl b/latency-aware/flappy/mean_rtx3090_s42_200ep_v1/demonstrations/conversion/expected/lerobot__val/meta/episodes.jsonl new file mode 100644 index 0000000000000000000000000000000000000000..874004118d9126e91a4f725c3be5daeb2a018a65 --- /dev/null +++ b/latency-aware/flappy/mean_rtx3090_s42_200ep_v1/demonstrations/conversion/expected/lerobot__val/meta/episodes.jsonl @@ -0,0 +1,20 @@ +{"episode_index": 0, "length": 18000} +{"episode_index": 1, "length": 18000} +{"episode_index": 2, "length": 18000} +{"episode_index": 3, "length": 18000} +{"episode_index": 4, "length": 18000} +{"episode_index": 5, "length": 18000} +{"episode_index": 6, "length": 18000} +{"episode_index": 7, "length": 18000} +{"episode_index": 8, "length": 18000} +{"episode_index": 9, "length": 18000} +{"episode_index": 10, "length": 18000} +{"episode_index": 11, "length": 18000} +{"episode_index": 12, "length": 18000} +{"episode_index": 13, "length": 18000} +{"episode_index": 14, "length": 18000} +{"episode_index": 15, "length": 18000} +{"episode_index": 16, "length": 18000} +{"episode_index": 17, "length": 18000} +{"episode_index": 18, "length": 18000} +{"episode_index": 19, "length": 18000} diff --git a/latency-aware/flappy/mean_rtx3090_s42_200ep_v1/demonstrations/conversion/expected/lerobot__val/meta/info.json b/latency-aware/flappy/mean_rtx3090_s42_200ep_v1/demonstrations/conversion/expected/lerobot__val/meta/info.json new file mode 100644 index 0000000000000000000000000000000000000000..74fd9be04402a3da116a334b50809b6f3eaab000 --- /dev/null +++ b/latency-aware/flappy/mean_rtx3090_s42_200ep_v1/demonstrations/conversion/expected/lerobot__val/meta/info.json @@ -0,0 +1,102 @@ +{ + "codebase_version": "v2.0", + "fps": 10.0, + "total_episodes": 20, + "total_frames": 360000, + "total_tasks": 1, + "total_videos": 0, + "total_chunks": 1, + "chunks_size": 1000, + "data_path": "data/chunk-{episode_chunk:03d}/episode_{episode_index:06d}.parquet", + "video_path": "videos/chunk-{episode_chunk:03d}/{video_key}/episode_{episode_index:06d}.mp4", + "features": { + "observation.image": { + "dtype": "image", + "shape": [ + 512, + 288, + 3 + ], + "names": [ + "height", + "width", + "channel" + ], + "video_info": { + "video.fps": 10.0 + } + }, + "observation.state": { + "dtype": "float32", + "shape": [ + 24 + ], + "names": [ + "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" + ] + }, + "action": { + "dtype": "float32", + "shape": [ + 2 + ], + "names": [ + "NOOP", + "FLAP" + ] + }, + "timestamp": { + "dtype": "float64", + "shape": [ + 1 + ] + }, + "episode_index": { + "dtype": "int64", + "shape": [ + 1 + ] + }, + "frame_index": { + "dtype": "int64", + "shape": [ + 1 + ] + }, + "task_index": { + "dtype": "int64", + "shape": [ + 1 + ] + }, + "latency": { + "dtype": "int64", + "shape": [ + 1 + ] + } + } +} \ No newline at end of file diff --git a/latency-aware/flappy/mean_rtx3090_s42_200ep_v1/demonstrations/conversion/expected/lerobot__val/meta/modality.json b/latency-aware/flappy/mean_rtx3090_s42_200ep_v1/demonstrations/conversion/expected/lerobot__val/meta/modality.json new file mode 100644 index 0000000000000000000000000000000000000000..9b611ad7557a5feabdc80e1c2f11a706a99bb45c --- /dev/null +++ b/latency-aware/flappy/mean_rtx3090_s42_200ep_v1/demonstrations/conversion/expected/lerobot__val/meta/modality.json @@ -0,0 +1,30 @@ +{ + "state": { + "game_state": { + "start": 0, + "end": 24, + "dtype": "float32", + "absolute": true, + "original_key": "observation.state" + } + }, + "action": { + "button": { + "start": 0, + "end": 2, + "dtype": "float32", + "absolute": true, + "original_key": "action" + } + }, + "video": { + "image": { + "original_key": "observation.image" + } + }, + "annotation": { + "human.action.task_description": { + "original_key": "task_index" + } + } +} diff --git a/latency-aware/flappy/mean_rtx3090_s42_200ep_v1/demonstrations/conversion/expected/lerobot__val/meta/tasks.jsonl b/latency-aware/flappy/mean_rtx3090_s42_200ep_v1/demonstrations/conversion/expected/lerobot__val/meta/tasks.jsonl new file mode 100644 index 0000000000000000000000000000000000000000..8a4acdb4011cc11914f414e2c5b7bf68740f8781 --- /dev/null +++ b/latency-aware/flappy/mean_rtx3090_s42_200ep_v1/demonstrations/conversion/expected/lerobot__val/meta/tasks.jsonl @@ -0,0 +1 @@ +{"task": "You are playing Flappy Bird. Pass through the pipe gaps and stay alive. Choose the action: NOOP, FLAP.", "task_index": 0} diff --git a/latency-aware/flappy/mean_rtx3090_s42_200ep_v1/demonstrations/raw/filter_report.json b/latency-aware/flappy/mean_rtx3090_s42_200ep_v1/demonstrations/raw/filter_report.json new file mode 100644 index 0000000000000000000000000000000000000000..1557d04124b19180c0f21901d73a7b9f973e8239 --- /dev/null +++ b/latency-aware/flappy/mean_rtx3090_s42_200ep_v1/demonstrations/raw/filter_report.json @@ -0,0 +1,12 @@ +{ + "accepted_episodes": 200, + "accepted_train_episodes": 180, + "accepted_val_episodes": 20, + "attempted_episodes": 200, + "preset": "none", + "rejected_episodes": 0, + "train_rows_after_row_filter": 3240000, + "train_rows_before_row_filter": 3240000, + "val_rows_after_row_filter": 360000, + "val_rows_before_row_filter": 360000 +} diff --git a/latency-aware/flappy/mean_rtx3090_s42_200ep_v1/demonstrations/raw/metadata.json b/latency-aware/flappy/mean_rtx3090_s42_200ep_v1/demonstrations/raw/metadata.json new file mode 100644 index 0000000000000000000000000000000000000000..294a3482612177be7ef0d108824f691c43fc604f --- /dev/null +++ b/latency-aware/flappy/mean_rtx3090_s42_200ep_v1/demonstrations/raw/metadata.json @@ -0,0 +1,363 @@ +{ + "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": { + "action_history_decisions": 8, + "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": { + "action_history_decisions": 8, + "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": 10.0, + "env_gpu_actions": true, + "env_gpu_observations": true, + "episode_metrics_path": "/data/ubuntu/lzj/mean-profiling/flappy/teacher/flappy_mean_3090_s42/episode_metrics.jsonl", + "eval_deterministic": true, + "eval_episodes": 100, + "eval_max_steps": 18000, + "eval_parallel_envs": 100, + "experiment": "flappy_mean_3090_s42", + "experiment_summaries_interval": 1, + "fixed_latency_ms": 75.87417450998383, + "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", + "include_transport_history": false, + "keep_checkpoints": 5, + "latency_profile_path": null, + "latency_profile_worker_slot": null, + "latency_seed": 42, + "latency_type": "fixed", + "learning_rate": 0.00025, + "lr_schedule": "linear_decay", + "mask_unadmitted_actor": false, + "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": 10.0, + "obs_resize": "84,84", + "ordering_policy": "latest_ready", + "output_dir": "/data/ubuntu/lzj/mean-profiling/flappy/teacher-log", + "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": 42, + "simulated_worker_capacity": 1, + "simulator": "gpu", + "stats_avg": 100, + "train_dir": "/data/ubuntu/lzj/mean-profiling/flappy/teacher", + "train_for_env_steps": 25000000, + "use_gpu_render": true, + "use_lidar": false, + "use_rnn": false, + "value_loss_coeff": 0.5, + "wandb_dir": "/data/ubuntu/lzj/mean-profiling/flappy/wandb", + "wandb_group": "mean-profiling-3090", + "wandb_job_type": "mean-profile-teacher", + "wandb_project": "starvla_tasks", + "wandb_tags": [ + "flappy", + "qwenoft", + "mean-profile" + ], + "wandb_unique_id": "mean3090-flappy-teacher-20260928", + "wandb_user": "dongqianyu99-zhejiang-university", + "with_wandb": true, + "worker_num_splits": 1, + "zero_worker_busy_bit": false + }, + "command_line": "Configured from pinned published config.json; see teacher-config-diff.json", + "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": 10.0, + "env_frameskip": 1, + "env_framestack": 1, + "env_gpu_actions": true, + "env_gpu_observations": true, + "episode_counter": false, + "episode_metrics_path": "/data/ubuntu/lzj/mean-profiling/flappy/teacher/flappy_mean_3090_s42/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": 18000, + "eval_parallel_envs": 100, + "eval_raw_reward": false, + "experiment": "flappy_mean_3090_s42", + "experiment_summaries_interval": 1, + "exploration_loss": "entropy", + "exploration_loss_coeff": 0.003, + "fasttd3_action_chunk_horizon": 1, + "fasttd3_actor_action_l2": 0.0, + "fasttd3_compile": true, + "fasttd3_replay_batch_size": 32768, + "fasttd3_replay_capacity": 6553600, + "fasttd3_train_for_optimizer_steps": 10000000000, + "fasttd3_transitions_per_update": 64, + "fasttd3_v_max": 250.0, + "fasttd3_v_min": -250.0, + "fixed_latency_ms": 75.87417450998383, + "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", + "include_transport_history": false, + "initial_model_path": null, + "initial_stddev": 1.0, + "keep_checkpoints": 5, + "kl_loss_coeff": 0.0, + "latency_profile_path": null, + "latency_profile_worker_slot": null, + "latency_seed": 42, + "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, + "lr_schedule_min_lr": 0.0, + "mask_unadmitted_actor": false, + "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": 10.0, + "obs_resize": "84,84", + "obs_scale": 255.0, + "obs_subtract_mean": 0.0, + "optimizer": "adam", + "ordering_policy": "latest_ready", + "output_dir": "/data/ubuntu/lzj/mean-profiling/flappy/teacher-log", + "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, + "simulated_worker_capacity": 1, + "simulator": "gpu", + "stats_avg": 100, + "std_latency_ms": null, + "summaries_use_frameskip": true, + "train_dir": "/data/ubuntu/lzj/mean-profiling/flappy/teacher", + "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_dir": "/data/ubuntu/lzj/mean-profiling/flappy/wandb", + "wandb_group": "mean-profiling-3090", + "wandb_job_type": "mean-profile-teacher", + "wandb_project": "starvla_tasks", + "wandb_tags": [ + "flappy", + "qwenoft", + "mean-profile" + ], + "wandb_unique_id": "mean3090-flappy-teacher-20260928", + "wandb_user": "dongqianyu99-zhejiang-university", + "with_pbt": false, + "with_vtrace": false, + "with_wandb": true, + "worker_num_splits": 1, + "zero_worker_busy_bit": false + }, + "checkpoint_experiment_dir": "/data/ubuntu/lzj/mean-profiling/flappy/restored-teacher/latency-aware/flappy/small-policy/sample-factory-qwenoft-rtx3090-mean-hist8-capacity1-s42-v1", + "checkpoint_path": "/data/ubuntu/lzj/mean-profiling/flappy/restored-teacher/latency-aware/flappy/small-policy/sample-factory-qwenoft-rtx3090-mean-hist8-capacity1-s42-v1/best_000048808_25034752_reward_2216.841.pth", + "checkpoint_train_step": 48808, + "config_source": "/data/ubuntu/lzj/mean-profiling/flappy/restored-teacher/latency-aware/flappy/small-policy/sample-factory-qwenoft-rtx3090-mean-hist8-capacity1-s42-v1/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": 3240000, + "train_rows_before_row_filter": 3240000, + "val_rows_after_row_filter": 360000, + "val_rows_before_row_filter": 360000 + }, + "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": 18000, + "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": 3600000, + "train_episodes": 180, + "train_rows": 3240000, + "val_episodes": 20, + "val_fraction": 0.1, + "val_rows": 360000 +} diff --git a/latency-aware/flappy/mean_rtx3090_s42_200ep_v1/demonstrations/raw/train.parquet b/latency-aware/flappy/mean_rtx3090_s42_200ep_v1/demonstrations/raw/train.parquet new file mode 100644 index 0000000000000000000000000000000000000000..9def0ae4a4d0af1abd88b8243fda3e1f82c21a05 --- /dev/null +++ b/latency-aware/flappy/mean_rtx3090_s42_200ep_v1/demonstrations/raw/train.parquet @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4843b4ba988dac95ec0a988e0ce5038c8459adf8314fafbfb11029313a2fbfd7 +size 36950461127 diff --git a/latency-aware/flappy/mean_rtx3090_s42_200ep_v1/demonstrations/raw/val.parquet b/latency-aware/flappy/mean_rtx3090_s42_200ep_v1/demonstrations/raw/val.parquet new file mode 100644 index 0000000000000000000000000000000000000000..40f70d905c45e9461ea439e74e0d5851d312a5f3 --- /dev/null +++ b/latency-aware/flappy/mean_rtx3090_s42_200ep_v1/demonstrations/raw/val.parquet @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3a6b5602fcbaa851ab0bbd0e2a0980ea235781d0fe0f2df83e93d35d732bde4b +size 4107439052 diff --git a/latency-aware/flappy/mean_rtx3090_s42_200ep_v1/demonstrations/validation/collection.complete.json b/latency-aware/flappy/mean_rtx3090_s42_200ep_v1/demonstrations/validation/collection.complete.json new file mode 100644 index 0000000000000000000000000000000000000000..b8819eaa2e9021f64ef937328c2125f1c0f0de42 --- /dev/null +++ b/latency-aware/flappy/mean_rtx3090_s42_200ep_v1/demonstrations/validation/collection.complete.json @@ -0,0 +1,15 @@ +{ + "task": "flappy", + "teacher_checkpoint": "/data/ubuntu/lzj/mean-profiling/flappy/restored-teacher/latency-aware/flappy/small-policy/sample-factory-qwenoft-rtx3090-mean-hist8-capacity1-s42-v1/best_000048808_25034752_reward_2216.841.pth", + "teacher_sha256": "7dd3923510bfe04612941b89bdf9d0e324bc51de12c999e5abc122f107ab2971", + "episodes": 200, + "train_episodes": 180, + "val_episodes": 20, + "train_rows": 3240000, + "val_rows": 360000, + "fixed_latency_ms": 75.87417450998383, + "action_horizon": 1, + "unique_seeds": 200, + "state_dim": 24, + "streaming_raw_rgb": true +} diff --git a/latency-aware/flappy/mean_rtx3090_s42_200ep_v1/demonstrations/validation/dataset.audit.json b/latency-aware/flappy/mean_rtx3090_s42_200ep_v1/demonstrations/validation/dataset.audit.json new file mode 100644 index 0000000000000000000000000000000000000000..2a1836cfe2ba08e16fa61201690dbd33147f0c70 --- /dev/null +++ b/latency-aware/flappy/mean_rtx3090_s42_200ep_v1/demonstrations/validation/dataset.audit.json @@ -0,0 +1,109 @@ +{ + "status": "passed", + "audit_utc": "2026-09-29T05:54:55.769025+00:00", + "scope": "CPU data audit; no model execution or independent task evaluation", + "raw_rows": { + "train": 3240000, + "validation": 360000 + }, + "episodes": { + "train": 180, + "validation": 20 + }, + "per_episode_steps": 18000, + "decision_steps": "0..17999 contiguous in every episode", + "canonical_seed_field": "seed", + "unique_seeds": 200, + "seed_range": [ + 42, + 241 + ], + "seed_mapping": "42 + original episode_idx", + "train_validation_seed_overlap": 0, + "state": { + "dimension": 24, + "encoding": "binary 0/1", + "history_decisions": 8, + "all_rows_checked": true, + "zero_reset_and_causal_shifts": "passed" + }, + "action": { + "dimension": 2, + "encoding": "one-hot", + "matches_action_id": "all rows" + }, + "latency": { + "configured_mean_ms": 75.87417450998383, + "raw_frames": 1, + "stored_quantized_ms": 100.0, + "env_fps": 10, + "obs_fps": 10, + "all_rows_checked": true + }, + "action_counts": { + "train": { + "noop": 2918864, + "flap": 321136 + }, + "validation": { + "noop": 324379, + "flap": 35621 + } + }, + "converted_episode_files_verified": 200, + "image_shape_hwc": [ + 512, + 288, + 3 + ], + "model_resize_wh": [ + 224, + 224 + ], + "paired_sample_frames": 30, + "sample_steps": [ + 0, + 1, + 8, + 9000, + 17999 + ], + "sample_train_converted_episodes": [ + 0, + 90, + 179 + ], + "sample_train_raw_episodes": [ + 0, + 98, + 199 + ], + "sample_validation_converted_episodes": [ + 0, + 10, + 19 + ], + "sample_validation_raw_episodes": [ + 3, + 115, + 196 + ], + "paired_sample_checks": [ + "PNG bytes", + "decoded pixels", + "state", + "action", + "reward", + "latency", + "timestamp", + "task", + "done" + ], + "modality_registry": "rl_games_flappy: FlappyDataConfig; declared keys match both split metadata and mixture", + "action_horizon": 1, + "limitations": "Raw numeric fields checked in full; raw-to-converted pixel/row identity checked on 30 sampled frames, not every image. Data loader was not instantiated, no GPU/model evaluation performed.", + "original_collect_wrapper_exit": 1, + "failure_cause": "Wrapper read temporary probe_state.json after canonical exporter cleanup", + "recollection_performed": false, + "finalize_only_exit": 0 +} diff --git a/latency-aware/flappy/mean_rtx3090_s42_200ep_v1/evidence/files.sha256.json b/latency-aware/flappy/mean_rtx3090_s42_200ep_v1/evidence/files.sha256.json new file mode 100644 index 0000000000000000000000000000000000000000..11564a6c4b5668fe1423be02681dbb6f5fe9fd59 --- /dev/null +++ b/latency-aware/flappy/mean_rtx3090_s42_200ep_v1/evidence/files.sha256.json @@ -0,0 +1,156 @@ +{ + "repo_id": "latency-sensitive-bench/Standard-Pipeline", + "path": "flappy/profile_latency/QwenOFT/mean_rtx3090_s42_200ep_v1", + "files": { + "flappy/README.md": { + "size": 521, + "sha256": "9871c695386e238fb8819a13f20c4143ae85917fd2e3f7974a9790f284b91a6c", + "git_blob_id": "f29fe44a2202e8949f0c16105256df32aa611dc6" + }, + "flappy/profile_latency/QwenOFT/mean_rtx3090_s42_200ep_v1/README.md": { + "size": 1953, + "sha256": "94c61619c7317d72613c807c3c509b67f46b72eeb1c35bf7b786fd7e6ece32ad", + "git_blob_id": "788eb8d4031d955cd83428b544a68456dacbcbec" + }, + "flappy/profile_latency/QwenOFT/mean_rtx3090_s42_200ep_v1/demonstrations/conversion/README.md": { + "size": 1671, + "sha256": "9d31e94394c7230ee4f9e631bce672090e34313f619d602dca7c38677fcb54d6", + "git_blob_id": "f8b6f62fdbc0bedac7a005746eabd4350b036a59" + }, + "flappy/profile_latency/QwenOFT/mean_rtx3090_s42_200ep_v1/demonstrations/conversion/convert_to_lerobot.py": { + "size": 396, + "sha256": "637e195c690addfd311d9d9df8dde995888f5af5616f5d5ecc5d47198b7b6c02", + "git_blob_id": "f0ed54d13fcef3f55732a5fb53a26328f2fc27d4" + }, + "flappy/profile_latency/QwenOFT/mean_rtx3090_s42_200ep_v1/demonstrations/conversion/expected/_generated_mixtures/lerobot.json": { + "size": 180, + "sha256": "4b789655fd60e90c5b60d485436e984a099d29c3c8065eb6967b16e29a1bc729", + "git_blob_id": "90240ec6f4201c9e5379b57ecae93e73500796a1" + }, + "flappy/profile_latency/QwenOFT/mean_rtx3090_s42_200ep_v1/demonstrations/conversion/expected/lerobot/latency_prompt_map.json": { + "size": 189, + "sha256": "e5bdad8c4747fb2e8101f5d7265e89e2c8794c573ee42bac764365a42b6bc3a4", + "git_blob_id": "465788c1cf15573a239a3146b6aca2bb8b5deaa4" + }, + "flappy/profile_latency/QwenOFT/mean_rtx3090_s42_200ep_v1/demonstrations/conversion/expected/lerobot/manifest.json": { + "size": 2295, + "sha256": "adee4e61a22ca183bbf8255b2924f20e052f591868bda708f5b64fd261942950", + "git_blob_id": "34f74342c9d2c900c37313a53a88d65e05e6125f" + }, + "flappy/profile_latency/QwenOFT/mean_rtx3090_s42_200ep_v1/demonstrations/conversion/expected/lerobot/meta/episodes.jsonl": { + "size": 7090, + "sha256": "bb22a5c6237f8ad1da7ed0daeea01afa433f811fb6c1254c1b0138fd1193a18e", + "git_blob_id": "d4cfdea0f4cb2c32438c8907c01c9226eab75699" + }, + "flappy/profile_latency/QwenOFT/mean_rtx3090_s42_200ep_v1/demonstrations/conversion/expected/lerobot/meta/info.json": { + "size": 2068, + "sha256": "d898f74d4dba22f86deb5a7d3d15a9a8ffe3d000caa60f239cd6524ee71b4a33", + "git_blob_id": "b0dd07442e6b4b29a7aec730386ed865e747375e" + }, + "flappy/profile_latency/QwenOFT/mean_rtx3090_s42_200ep_v1/demonstrations/conversion/expected/lerobot/meta/modality.json": { + "size": 514, + "sha256": "46999cb6f26a99b66d2dcbfcda47d80abf24bdee3bf0a09cec9d365d1cc338c6", + "git_blob_id": "9b611ad7557a5feabdc80e1c2f11a706a99bb45c" + }, + "flappy/profile_latency/QwenOFT/mean_rtx3090_s42_200ep_v1/demonstrations/conversion/expected/lerobot/meta/tasks.jsonl": { + "size": 132, + "sha256": "7b197930d4e3e216bc3abc3db9fa6ee9dc58c4e0aeeb072a75f753da27c0b0ec", + "git_blob_id": "8a4acdb4011cc11914f414e2c5b7bf68740f8781" + }, + "flappy/profile_latency/QwenOFT/mean_rtx3090_s42_200ep_v1/demonstrations/conversion/expected/lerobot__val/manifest.json": { + "size": 2231, + "sha256": "ea638865557b20c77fb60ab4863ed2a20c021bc24c12daf83f11b77153f6ce8c", + "git_blob_id": "2c8fd2d341f04397e8243fb451f8ad74dfbb3a0d" + }, + "flappy/profile_latency/QwenOFT/mean_rtx3090_s42_200ep_v1/demonstrations/conversion/expected/lerobot__val/meta/episodes.jsonl": { + "size": 770, + "sha256": "0471962d22560b3af7cd16ae468c67ee9fc7937b2414285b4b91444200412148", + "git_blob_id": "874004118d9126e91a4f725c3be5daeb2a018a65" + }, + "flappy/profile_latency/QwenOFT/mean_rtx3090_s42_200ep_v1/demonstrations/conversion/expected/lerobot__val/meta/info.json": { + "size": 2066, + "sha256": "2cc62d8296e3bd89eb2c8a93a0c8cbdb5402837106de41ba5310455376bd0105", + "git_blob_id": "74fd9be04402a3da116a334b50809b6f3eaab000" + }, + "flappy/profile_latency/QwenOFT/mean_rtx3090_s42_200ep_v1/demonstrations/conversion/expected/lerobot__val/meta/modality.json": { + "size": 514, + "sha256": "46999cb6f26a99b66d2dcbfcda47d80abf24bdee3bf0a09cec9d365d1cc338c6", + "git_blob_id": "9b611ad7557a5feabdc80e1c2f11a706a99bb45c" + }, + "flappy/profile_latency/QwenOFT/mean_rtx3090_s42_200ep_v1/demonstrations/conversion/expected/lerobot__val/meta/tasks.jsonl": { + "size": 132, + "sha256": "7b197930d4e3e216bc3abc3db9fa6ee9dc58c4e0aeeb072a75f753da27c0b0ec", + "git_blob_id": "8a4acdb4011cc11914f414e2c5b7bf68740f8781" + }, + "flappy/profile_latency/QwenOFT/mean_rtx3090_s42_200ep_v1/demonstrations/raw/filter_report.json": { + "size": 335, + "sha256": "2c9d4f33c92cfc6d5b078ceef848608f67d11f4ca9e107cdb7a38670c2cec548", + "git_blob_id": "1557d04124b19180c0f21901d73a7b9f973e8239" + }, + "flappy/profile_latency/QwenOFT/mean_rtx3090_s42_200ep_v1/demonstrations/raw/metadata.json": { + "size": 11795, + "sha256": "898161f385ec7f8fa0ca0ca4c64d3023d8a0fa4e58bcc9e4a49254582b242f3d", + "git_blob_id": "294a3482612177be7ef0d108824f691c43fc604f" + }, + "flappy/profile_latency/QwenOFT/mean_rtx3090_s42_200ep_v1/demonstrations/raw/train.parquet": { + "size": 36950461127, + "sha256": "4843b4ba988dac95ec0a988e0ce5038c8459adf8314fafbfb11029313a2fbfd7", + "git_blob_id": null + }, + "flappy/profile_latency/QwenOFT/mean_rtx3090_s42_200ep_v1/demonstrations/raw/val.parquet": { + "size": 4107439052, + "sha256": "3a6b5602fcbaa851ab0bbd0e2a0980ea235781d0fe0f2df83e93d35d732bde4b", + "git_blob_id": null + }, + "flappy/profile_latency/QwenOFT/mean_rtx3090_s42_200ep_v1/demonstrations/validation/collection.complete.json": { + "size": 579, + "sha256": "5db0a297a1ccf00e86940f527467bd4e1c33b208c3af181632065b93d82c5339", + "git_blob_id": "b8819eaa2e9021f64ef937328c2125f1c0f0de42" + }, + "flappy/profile_latency/QwenOFT/mean_rtx3090_s42_200ep_v1/demonstrations/validation/dataset.audit.json": { + "size": 2385, + "sha256": "97e8dfe029598ec854cf767c54fd94d139305bfe2a0d4814fd4d1f435ca1ff2e", + "git_blob_id": "2a1836cfe2ba08e16fa61201690dbd33147f0c70" + }, + "flappy/profile_latency/QwenOFT/mean_rtx3090_s42_200ep_v1/evidence/provenance/collect_sf.py": { + "size": 2891, + "sha256": "a146b66b548a9f1589584f499583f0804387f50b4a3e06ff61402853f988823e", + "git_blob_id": "dd6f9fe017ddc2fbe4e0f0a06e4a122eb1712f31" + }, + "flappy/profile_latency/QwenOFT/mean_rtx3090_s42_200ep_v1/evidence/provenance/collection-command.json": { + "size": 792, + "sha256": "74d70e305ffca4d69d47b43cf284a699765241425e04e1c051f20ca6a58d2a8c", + "git_blob_id": "107dd860f750b35c8267e6d93899cf213dca33fa" + }, + "flappy/profile_latency/QwenOFT/mean_rtx3090_s42_200ep_v1/evidence/provenance/source.json": { + "size": 2609, + "sha256": "881731dd86e136ec501c5935b013f520a20edfd73e73e4ccc567da08643bc390", + "git_blob_id": "642400faac54314812fdaf1b1a840146783117ae" + }, + "flappy/profile_latency/QwenOFT/mean_rtx3090_s42_200ep_v1/evidence/provenance/starvla-data_config.source.py": { + "size": 6422, + "sha256": "30bf7c0648642364c7162466e0cdbe42075efe78b393141b9cd6f1be2cd9cf8a", + "git_blob_id": "6d76aebc2dca09af28fc6304ad45f6dd19499952" + }, + "flappy/profile_latency/QwenOFT/mean_rtx3090_s42_200ep_v1/evidence/provenance/starvla_lerobot.source.py": { + "size": 42762, + "sha256": "c3aae215af7c7f04051d7a2731a2007727f7ac3426fbc8d2aad2f6a276ae842b", + "git_blob_id": "fba03402d2c26da376833bead62f0b17d4fed9cd" + }, + "flappy/profile_latency/QwenOFT/mean_rtx3090_s42_200ep_v1/evidence/provenance/stream-raw-rgb.patch": { + "size": 5619, + "sha256": "47a877d1c7b4dabec6dc5cf573848bdec8690bc7fd323a9272de39d7914fec80", + "git_blob_id": "d3af4377165f253a9f090b55019c6b800d7bdc9e" + }, + "flappy/profile_latency/QwenOFT/mean_rtx3090_s42_200ep_v1/evidence/provenance/vla.training-config.reference.yaml": { + "size": 6574, + "sha256": "edcbc8fc1e86209fe50533d1dbbb194b585eaf378e814b1fe300b3820bc7c48b", + "git_blob_id": "6347967a8d8127a273bdec69b222541bd32a2296" + }, + "flappy/profile_latency/QwenOFT/mean_rtx3090_s42_200ep_v1/profiling/source/profile.json": { + "size": 2140, + "sha256": "c266cba7ebac05c7e37bc83f1f16fe4b27dab97942ff43290e6c41514f6e39fc", + "git_blob_id": "1dc1af8afd0740a8e69efc4ad41b1e3173bdf3fc" + } + } +} diff --git a/latency-aware/flappy/mean_rtx3090_s42_200ep_v1/evidence/provenance/collect_sf.py b/latency-aware/flappy/mean_rtx3090_s42_200ep_v1/evidence/provenance/collect_sf.py new file mode 100644 index 0000000000000000000000000000000000000000..dd6f9fe017ddc2fbe4e0f0a06e4a122eb1712f31 --- /dev/null +++ b/latency-aware/flappy/mean_rtx3090_s42_200ep_v1/evidence/provenance/collect_sf.py @@ -0,0 +1,25 @@ +from pathlib import Path +import argparse,hashlib,json,subprocess,sys +import pyarrow.parquet as pq +from latency_bench.data.exporters.starvla_lerobot import convert_canonical_rollout_to_lerobot +ROOT=Path('/data/ubuntu/lzj/mean-profiling');REPO=Path('/data/ubuntu/lzj/latency-sensitive-bench') +BUDGETS={'flappy':(200,18000,256,256,'gpu'),'deadly_corridor':(8500,3600,16,8,'cpu'),'ant':(100,1000,12,8,'gpu')} +parser=argparse.ArgumentParser();parser.add_argument('task',choices=BUDGETS);parser.add_argument('--finalize-only',action='store_true');args=parser.parse_args() +task=args.task;out=ROOT/task;cfg=json.loads((out/'teacher.flat.json').read_text());checkpoint=Path(json.loads((out/'teacher.upload.json').read_text())['local_checkpoint']) +episodes,horizon,probe_envs,replay_envs,device=BUDGETS[task];raw=out/'data/raw' +command=[sys.executable,str(REPO/'scripts/rollout_data/export_sf_teacher_dataset.py'),'--checkpoint-root',str(checkpoint),'--output-dir',str(raw),'--max-episodes',str(episodes),'--max-steps-per-episode',str(horizon),'--num-envs',str(probe_envs),'--replay-num-envs',str(replay_envs),'--device',device,'--val-fraction','0.1','--split-seed','0','--context-window','1','--filter-preset','none','--deterministic','--image-writer-workers','32' if device=='gpu' else '8'] +if task=='ant':command.extend(['--episode-return-gt','2500','--export-image-size','224,224']) +else:command.extend(['--max-attempts',str(episodes)]) +(out/'collection-command.json').write_text(json.dumps(command,indent=2)+'\n') +if not args.finalize_only: + subprocess.run(command,cwd=REPO,check=True) + convert_canonical_rollout_to_lerobot(raw,out/'data/lerobot',action_carrier='native') +manifest=json.loads((out/'data/lerobot/manifest.json').read_text());meta=json.loads((raw/'metadata.json').read_text()) +assert meta['total_episodes']==episodes and (meta['env_fps'],meta['obs_fps'])==(cfg['env_fps'],cfg['obs_fps']) +assert manifest['state_dim']=={'flappy':24,'deadly_corridor':120,'ant':27}[task] +train_seeds=set(pq.read_table(raw/'train.parquet',columns=['seed']).column('seed').unique().to_pylist()) +val_seeds=set(pq.read_table(raw/'val.parquet',columns=['seed']).column('seed').unique().to_pylist()) +assert len(train_seeds)==meta['train_episodes'] and len(val_seeds)==meta['val_episodes'] and train_seeds.isdisjoint(val_seeds) +seeds=train_seeds | val_seeds +report={'task':task,'teacher_checkpoint':str(checkpoint),'teacher_sha256':hashlib.sha256(checkpoint.read_bytes()).hexdigest(),'episodes':episodes,'train_episodes':meta['train_episodes'],'val_episodes':meta['val_episodes'],'train_rows':meta['train_rows'],'val_rows':meta['val_rows'],'fixed_latency_ms':cfg['fixed_latency_ms'],'action_horizon':1,'unique_seeds':len(seeds),'state_dim':manifest['state_dim'],'streaming_raw_rgb':True} +(out/'data/collection.complete.json').write_text(json.dumps(report,indent=2)+'\n');print(json.dumps(report),flush=True) diff --git a/latency-aware/flappy/mean_rtx3090_s42_200ep_v1/evidence/provenance/collection-command.json b/latency-aware/flappy/mean_rtx3090_s42_200ep_v1/evidence/provenance/collection-command.json new file mode 100644 index 0000000000000000000000000000000000000000..107dd860f750b35c8267e6d93899cf213dca33fa --- /dev/null +++ b/latency-aware/flappy/mean_rtx3090_s42_200ep_v1/evidence/provenance/collection-command.json @@ -0,0 +1,31 @@ +[ + "/data/ubuntu/lzj/conda/envs/qwenoft/bin/python", + "/data/ubuntu/lzj/latency-sensitive-bench/scripts/rollout_data/export_sf_teacher_dataset.py", + "--checkpoint-root", + "/data/ubuntu/lzj/mean-profiling/flappy/restored-teacher/latency-aware/flappy/small-policy/sample-factory-qwenoft-rtx3090-mean-hist8-capacity1-s42-v1/best_000048808_25034752_reward_2216.841.pth", + "--output-dir", + "/data/ubuntu/lzj/mean-profiling/flappy/data/raw", + "--max-episodes", + "200", + "--max-steps-per-episode", + "18000", + "--num-envs", + "256", + "--replay-num-envs", + "256", + "--device", + "gpu", + "--val-fraction", + "0.1", + "--split-seed", + "0", + "--context-window", + "1", + "--filter-preset", + "none", + "--deterministic", + "--image-writer-workers", + "32", + "--max-attempts", + "200" +] diff --git a/latency-aware/flappy/mean_rtx3090_s42_200ep_v1/evidence/provenance/source.json b/latency-aware/flappy/mean_rtx3090_s42_200ep_v1/evidence/provenance/source.json new file mode 100644 index 0000000000000000000000000000000000000000..642400faac54314812fdaf1b1a840146783117ae --- /dev/null +++ b/latency-aware/flappy/mean_rtx3090_s42_200ep_v1/evidence/provenance/source.json @@ -0,0 +1,60 @@ +{ + "repository_archive_base_revision": "ff90f1b6c46a6eaae55bbf341762f03a2be6e013", + "condition": "fixed mean latency derived from the cited RTX3090 profile; no temporal/distribution sampling", + "fixed_latency_ms": 75.87417450998383, + "env_fps": 10, + "obs_fps": 10, + "quantized_latency_raw_frames": 1, + "stored_quantized_latency_ms": 100.0, + "teacher": { + "repo_id": "latency-sensitive-bench/benchmark-models", + "revision": "3ca75d71b55ff42889a4f058d6ad90b2e9b3668f", + "path": "latency-aware/flappy/small-policy/sample-factory-qwenoft-rtx3090-mean-hist8-capacity1-s42-v1", + "checkpoint_filename": "best_000048808_25034752_reward_2216.841.pth", + "checkpoint_sha256": "7dd3923510bfe04612941b89bdf9d0e324bc51de12c999e5abc122f107ab2971", + "env_steps": 25034752 + }, + "profile": { + "repo_id": "latency-sensitive-bench/profiles", + "revision": "0de3f61f3635e7e706487c6cae62a4897bf636d5", + "path_in_repo": "profiles/openvla/1x-rtx3090/flappy/instance_a5037b165aa0cedc/profile.json", + "model_id": "openvla", + "gpu_class": "1x-rtx3090", + "source_profile_env_fps": 30 + }, + "code": { + "latency_sensitive_bench_revision": "c3c6a39365a151e9b7a5e215452fd64e957c2b29", + "starVLA_revision": "ccca13c5177fe3d3c884b6e2de4965d916016649", + "streaming_patch_sha256": "47a877d1c7b4dabec6dc5cf573848bdec8690bc7fd323a9272de39d7914fec80", + "converter_sha256": "c3aae215af7c7f04051d7a2731a2007727f7ac3426fbc8d2aad2f6a276ae842b", + "registry_sha256": "30bf7c0648642364c7162466e0cdbe42075efe78b393141b9cd6f1be2cd9cf8a" + }, + "conversion": { + "entrypoint": "latency_bench.data.exporters.starvla_lerobot.convert_canonical_rollout_to_lerobot", + "action_carrier": "native", + "force": false, + "normalize_state": true, + "write_starvla_handoff": true, + "format": "starvla_lerobot_v2_decision_v1", + "history_decisions": 8, + "state_dim": 24, + "action_dim": 2, + "action_horizon": 1, + "python": "3.10.21", + "packages": { + "huggingface-hub": "0.36.2", + "pyarrow": "25.0.1", + "numpy": "1.26.4", + "pandas": "2.3.3", + "datasets": "4.8.5", + "pillow": "12.1.1", + "sample-factory": "2.1.1", + "torch": "2.11.0+cu128", + "transformers": "4.57.0", + "deepspeed": "0.16.9", + "flash-attn": "2.8.3+cu12torch2.11cxx11abiTRUE" + } + }, + "distribution": "Raw canonical files only. Converted image payloads remain on the training host; conversion expected metadata is archived without duplicate image Parquet.", + "validation_scope": "CPU data validation only; no task/model performance evaluation is claimed." +} diff --git a/latency-aware/flappy/mean_rtx3090_s42_200ep_v1/evidence/provenance/starvla-data_config.source.py b/latency-aware/flappy/mean_rtx3090_s42_200ep_v1/evidence/provenance/starvla-data_config.source.py new file mode 100644 index 0000000000000000000000000000000000000000..6d76aebc2dca09af28fc6304ad45f6dd19499952 --- /dev/null +++ b/latency-aware/flappy/mean_rtx3090_s42_200ep_v1/evidence/provenance/starvla-data_config.source.py @@ -0,0 +1,155 @@ +"""RL-games data registry for StarVLA LeRobot datasets.""" + +from starVLA.dataloader.gr00t_lerobot.embodiment_tags import EmbodimentTag + + +class FlappyDataConfig: + embodiment_tag = EmbodimentTag.NEW_EMBODIMENT + video_keys = ["video.image"] + state_keys = ["state.game_state"] + action_keys = ["action.button"] + language_keys = ["annotation.human.action.task_description"] + observation_indices = [0] + action_indices = [0] + + def modality_config(self): + from starVLA.dataloader.gr00t_lerobot.datasets import ModalityConfig + + return { + "video": ModalityConfig(delta_indices=self.observation_indices, modality_keys=self.video_keys), + "state": ModalityConfig(delta_indices=self.observation_indices, modality_keys=self.state_keys), + "action": ModalityConfig(delta_indices=self.action_indices, modality_keys=self.action_keys), + "language": ModalityConfig(delta_indices=self.observation_indices, modality_keys=self.language_keys), + } + + def transform(self): + from starVLA.dataloader.gr00t_lerobot.transform.base import ComposedModalityTransform + from starVLA.dataloader.gr00t_lerobot.transform.state_action import StateActionToTensor + + return ComposedModalityTransform(transforms=[ + StateActionToTensor(apply_to=self.state_keys), + StateActionToTensor(apply_to=self.action_keys), + ]) + + +class DemonAttackDataConfig(FlappyDataConfig): + pass + + +class DefendTheLineDataConfig(FlappyDataConfig): + pass + + +class DeadlyCorridorDataConfig(FlappyDataConfig): + pass + + +class AsterixDataConfig(FlappyDataConfig): + pass + + +class AtlantisDataConfig(FlappyDataConfig): + pass + + +class GymnasiumDataConfig(FlappyDataConfig): + language_keys = ["annotation.human.task_description"] + + +class GymnasiumNativeDataConfig(GymnasiumDataConfig): + """Configure native Gymnasium state and action vectors.""" + + state_keys = ["state.native"] + action_keys = ["action.native"] + + def transform(self): + from starVLA.dataloader.gr00t_lerobot.transform.base import ComposedModalityTransform + from starVLA.dataloader.gr00t_lerobot.transform.state_action import ( + StateActionToTensor, + StateActionTransform, + ) + + keys = [*self.state_keys, *self.action_keys] + return ComposedModalityTransform( + transforms=[ + StateActionToTensor(apply_to=keys), + StateActionTransform( + apply_to=keys, + normalization_modes={ + "state.native": "q99", + "action.native": "q99", + }, + ), + ] + ) + + +ROBOT_TYPE_CONFIG_MAP = { + "rl_games_flappy": FlappyDataConfig(), + "rl_games_demon_attack": DemonAttackDataConfig(), + "rl_games_defend_the_line": DefendTheLineDataConfig(), + "rl_games_deadly_corridor": DeadlyCorridorDataConfig(), + "rl_games_asterix": AsterixDataConfig(), + "rl_games_atlantis": AtlantisDataConfig(), + "rl_games_gymnasium": GymnasiumDataConfig(), + "rl_games_gymnasium_discrete": GymnasiumDataConfig(), + "rl_games_gymnasium_native": GymnasiumNativeDataConfig(), +} + +ROBOT_TYPE_TO_EMBODIMENT_TAG = { + "rl_games_flappy": EmbodimentTag.NEW_EMBODIMENT, + "rl_games_demon_attack": EmbodimentTag.NEW_EMBODIMENT, + "rl_games_defend_the_line": EmbodimentTag.NEW_EMBODIMENT, + "rl_games_deadly_corridor": EmbodimentTag.NEW_EMBODIMENT, + "rl_games_asterix": EmbodimentTag.NEW_EMBODIMENT, + "rl_games_atlantis": EmbodimentTag.NEW_EMBODIMENT, + "rl_games_gymnasium": EmbodimentTag.NEW_EMBODIMENT, + "rl_games_gymnasium_discrete": EmbodimentTag.NEW_EMBODIMENT, + "rl_games_gymnasium_native": EmbodimentTag.NEW_EMBODIMENT, +} + +DATASET_NAMED_MIXTURES = { + "flappy_train": [("flappy_train", 1.0, "rl_games_flappy")], + "flappy_train__bridge": [("flappy_train__bridge", 1.0, "rl_games_flappy")], + "flappy_mixed_latency_train": [("flappy_mixed_latency_train", 1.0, "rl_games_flappy")], + "flappy_mixed_latency_train__bridge": [ + ("flappy_mixed_latency_train__bridge", 1.0, "rl_games_flappy") + ], + "demon_attack_train": [("demon_attack_train", 1.0, "rl_games_demon_attack")], + "demon_attack_train__bridge": [("demon_attack_train__bridge", 1.0, "rl_games_demon_attack")], + "demon_attack_mixed_latency_train": [("demon_attack_mixed_latency_train", 1.0, "rl_games_demon_attack")], + "demon_attack_mixed_latency_train__bridge": [ + ("demon_attack_mixed_latency_train__bridge", 1.0, "rl_games_demon_attack") + ], + "defend_the_line_train": [("defend_the_line_train", 1.0, "rl_games_defend_the_line")], + "defend_the_line_train__bridge": [ + ("defend_the_line_train__bridge", 1.0, "rl_games_defend_the_line") + ], + "defend_the_line_mixed_latency_train": [ + ("defend_the_line_mixed_latency_train", 1.0, "rl_games_defend_the_line") + ], + "defend_the_line_mixed_latency_train__bridge": [ + ("defend_the_line_mixed_latency_train__bridge", 1.0, "rl_games_defend_the_line") + ], + "deadly_corridor_train": [("deadly_corridor_train", 1.0, "rl_games_deadly_corridor")], + "deadly_corridor_train__bridge": [("deadly_corridor_train__bridge", 1.0, "rl_games_deadly_corridor")], + "deadly_corridor_mixed_latency_train": [ + ("deadly_corridor_mixed_latency_train", 1.0, "rl_games_deadly_corridor") + ], + "deadly_corridor_mixed_latency_train__bridge": [ + ("deadly_corridor_mixed_latency_train__bridge", 1.0, "rl_games_deadly_corridor") + ], + "asterix_train": [("asterix_train", 1.0, "rl_games_asterix")], + "asterix_train__bridge": [("asterix_train__bridge", 1.0, "rl_games_asterix")], + "asterix_mixed_latency_train": [("asterix_mixed_latency_train", 1.0, "rl_games_asterix")], + "asterix_mixed_latency_train__bridge": [ + ("asterix_mixed_latency_train__bridge", 1.0, "rl_games_asterix") + ], + "atlantis_train": [("atlantis_train", 1.0, "rl_games_atlantis")], + "atlantis_train__bridge": [("atlantis_train__bridge", 1.0, "rl_games_atlantis")], + "atlantis_mixed_latency_train": [("atlantis_mixed_latency_train", 1.0, "rl_games_atlantis")], + "atlantis_mixed_latency_train__bridge": [ + ("atlantis_mixed_latency_train__bridge", 1.0, "rl_games_atlantis") + ], + "h1hand_balance_hard": [("h1hand_balance_hard", 1.0, "rl_games_gymnasium_native")], +} diff --git a/latency-aware/flappy/mean_rtx3090_s42_200ep_v1/evidence/provenance/starvla_lerobot.source.py b/latency-aware/flappy/mean_rtx3090_s42_200ep_v1/evidence/provenance/starvla_lerobot.source.py new file mode 100644 index 0000000000000000000000000000000000000000..fba03402d2c26da376833bead62f0b17d4fed9cd --- /dev/null +++ b/latency-aware/flappy/mean_rtx3090_s42_200ep_v1/evidence/provenance/starvla_lerobot.source.py @@ -0,0 +1,1141 @@ +"""Convert canonical rollout exports to the StarVLA LeRobot image format.""" + +from __future__ import annotations + +import io +import json +import shutil +from pathlib import Path +from typing import Any, Mapping + +import numpy as np +import pyarrow as pa +import pyarrow.parquet as pq +from PIL import Image + +from latency_bench.utils.io import write_jsonl + +from latency_bench.data.action_layouts import ( + ACTION_LAYOUT_GYMNASIUM_CONTINUOUS_V1, + ACTION_LAYOUT_GYMNASIUM_DISCRETE_V1, + ACTION_LAYOUT_G1_SONIC_ABSOLUTE_JOINT_TARGET_V1, + action_spec_for_env, +) +from latency_bench.data.filters import density_weights_clean_v1 +from latency_bench.data.state_normalization import min_max_normalize_state + +STATE_DIM = 1 +BRIDGE_ACTION_DIM = 7 +LE_ROBOT_CHUNK_SIZE = 1000 +_ROBOT_TYPES_BY_ENV = { + "flappy": "rl_games_flappy", + "demon_attack": "rl_games_demon_attack", + "deadly_corridor": "rl_games_deadly_corridor", + "atlantis": "rl_games_atlantis", + "asterix": "rl_games_asterix", +} +_HISTORY_STATE_DIMS = {"flappy": 24, "demon_attack": 88, "deadly_corridor": 120, "atlantis": 56, "asterix": 136} +_DEADLY_MULTIBINARY_LABELS = [ + "MOVE_FORWARD", "MOVE_BACKWARD", "MOVE_LEFT", "MOVE_RIGHT", + "TURN_LEFT", "TURN_RIGHT", "ATTACK", +] + + +def write_starvla_modality( + dataset_dir: Path | str, + state_dim: int, + action_dim: int, + *, + action_key: str = "action", +) -> Path: + path = Path(dataset_dir) / "meta" / "modality.json" + path.parent.mkdir(parents=True, exist_ok=True) + path.write_text( + json.dumps( + { + "state": { + "state": { + "start": 0, + "end": state_dim, + "dtype": "float32", + "absolute": True, + "original_key": "observation.state", + } + }, + "action": { + action_key: { + "start": 0, + "end": action_dim, + "dtype": "float32", + "absolute": True, + "original_key": action_key, + } + }, + "video": {"ego_view": {"original_key": "observation.image"}}, + "annotation": { + "human.task_description": {"original_key": "task_index"} + }, + }, + indent=2, + ) + + "\n", + encoding="utf-8", + ) + return path + + +def _action_carrier_spec( + action_labels: list[str], + action_carrier: str, +) -> tuple[list[str], int, int, list[str]]: + if action_carrier == "native": + return action_labels, len(action_labels), STATE_DIM, ["state"] + if action_carrier != "bridge": + raise ValueError(f"Unsupported action_carrier={action_carrier!r}") + if len(action_labels) > BRIDGE_ACTION_DIM: + raise ValueError( + f"bridge action carrier supports at most {BRIDGE_ACTION_DIM} actions, " + f"got {len(action_labels)}" + ) + carrier_labels = [ + *action_labels, + *( + f"BRIDGE_PAD_{index}" + for index in range(len(action_labels), BRIDGE_ACTION_DIM) + ), + ] + state_labels = [f"BRIDGE_STATE_{index}" for index in range(BRIDGE_ACTION_DIM)] + return carrier_labels, len(action_labels), BRIDGE_ACTION_DIM, state_labels + + +def _robot_type_for_env(env_name: str, integration_name: str) -> str: + if integration_name == "gymnasium": + return "rl_games_gymnasium" + return _ROBOT_TYPES_BY_ENV[env_name] + + +def _write_starvla_handoff_files( + output_dir: Path, + val_output_dir: Path, + *, + env_name: str, + integration_name: str, + rows: list[dict[str, Any]], +) -> tuple[Path, Path]: + prompt_map = { + str(int(row["latency_raw_frames"])): { + "prompt": row["prompt"], + "latency_raw_frames": int(row["latency_raw_frames"]), + "latency_ms": float(row["latency_ms"]), + } + for row in rows + } + prompt_map_path = output_dir / "latency_prompt_map.json" + prompt_map_path.write_text(json.dumps(prompt_map, indent=2), encoding="utf-8") + + mixture_path = output_dir.parent / "_generated_mixtures" / f"{output_dir.name}.json" + mixture_path.parent.mkdir(parents=True, exist_ok=True) + robot_type = _robot_type_for_env(env_name, integration_name) + mixture_path.write_text( + json.dumps( + { + output_dir.name: [[output_dir.name, 1.0, robot_type]], + val_output_dir.name: [[val_output_dir.name, 1.0, robot_type]], + }, + indent=2, + ), + encoding="utf-8", + ) + return prompt_map_path, mixture_path + + +def _png_bytes(image: Any) -> bytes: + """Serialize a rollout image value to PNG bytes. + + Args: + image: Hugging Face image struct, PIL image, or array-like image. + + Returns: + RGB PNG bytes. + """ + if isinstance(image, Mapping): + if image["bytes"] is not None: + return bytes(image["bytes"]) + with Path(image["path"]).open("rb") as handle: + return handle.read() + if not isinstance(image, Image.Image): + image = Image.fromarray(np.asarray(image)) + image = image.convert("RGB") + buffer = io.BytesIO() + image.save(buffer, format="PNG") + return buffer.getvalue() + + +def _load_rows(raw_export_dir: Path, split: str) -> list[dict[str, Any]]: + """Load canonical rollout rows from one parquet split. + + Args: + raw_export_dir: Directory containing ``train.parquet`` and ``val.parquet``. + split: Split name to read. + + Returns: + Rows as Python dictionaries. + """ + path = raw_export_dir / f"{split}.parquet" + # Bound nested image buffers per Arrow batch; the converter still owns all rows. + return [ + row + for batch in pq.ParquetFile(path).iter_batches(batch_size=1024) + for row in batch.to_pylist() + ] + + +def _episode_groups(rows: list[dict[str, Any]]) -> list[list[dict[str, Any]]]: + """Group canonical rows by episode and order by decision step. + + Args: + rows: Canonical rollout rows from one split. + + Returns: + Episode row groups sorted by episode index. + """ + grouped: dict[int, list[dict[str, Any]]] = {} + for row in rows: + grouped.setdefault(int(row["episode_idx"]), []).append(row) + return [ + sorted(episode_rows, key=lambda row: int(row["decision_step"])) + for _, episode_rows in sorted(grouped.items()) + ] + + +def _legacy_row_action_vector(row: dict[str, Any], *, action_dim: int) -> list[float]: + if "action" in row and row["action"] is not None: + return [float(value) for value in row["action"]] + if "action_id" not in row: + raise KeyError(f"row is missing action fields; available columns: {sorted(row)}") + action_id = int(row["action_id"]) + if action_id < 0 or action_id >= action_dim: + raise ValueError( + f"action_id={action_id} is outside action range [0, {action_dim - 1}]" + ) + return [1.0 if index == action_id else 0.0 for index in range(action_dim)] + + +def _write_legacy_episode( + path: Path, + rows: list[dict[str, Any]], + *, + action_dim: int, +) -> None: + path.parent.mkdir(parents=True, exist_ok=True) + table = pa.table( + { + "observation.image": pa.array( + [{"bytes": row["image_bytes"], "path": None} for row in rows], + type=pa.struct([("bytes", pa.binary()), ("path", pa.string())]), + ), + "observation.state": pa.array( + [[0.0] * STATE_DIM for _ in rows], + type=pa.list_(pa.float32(), STATE_DIM), + ), + "action": pa.array( + [row["action"] for row in rows], + type=pa.list_(pa.float32(), action_dim), + ), + "timestamp": pa.array( + [row["timestamp"] for row in rows], type=pa.float64() + ), + "episode_index": pa.array( + [row["episode_index"] for row in rows], type=pa.int64() + ), + "frame_index": pa.array( + [row["frame_index"] for row in rows], type=pa.int64() + ), + "task_index": pa.array( + [row["task_index"] for row in rows], type=pa.int64() + ), + "done": pa.array([row["done"] for row in rows], type=pa.bool_()), + "reward": pa.array([row["reward"] for row in rows], type=pa.float32()), + "density_weight": pa.array( + [row.get("density_weight", 1.0) for row in rows], + type=pa.float32(), + ), + } + ) + pq.write_table(table, path) + + +def _write_legacy_metadata( + dataset_dir: Path, + *, + fps: float, + action_labels: list[str], + episode_lengths: list[int], + task_prompts: list[str], +) -> None: + meta_dir = dataset_dir / "meta" + meta_dir.mkdir(parents=True, exist_ok=True) + action_dim = len(action_labels) + (meta_dir / "info.json").write_text( + json.dumps( + { + "codebase_version": "v2.0", + "fps": fps, + "features": { + "observation.image": {"dtype": "image"}, + "observation.state": { + "dtype": "float32", + "shape": [STATE_DIM], + }, + "action": { + "dtype": "float32", + "shape": [action_dim], + "names": action_labels, + }, + "timestamp": {"dtype": "float64", "shape": [1]}, + "episode_index": {"dtype": "int64", "shape": [1]}, + "frame_index": {"dtype": "int64", "shape": [1]}, + "task_index": {"dtype": "int64", "shape": [1]}, + }, + }, + indent=2, + ), + encoding="utf-8", + ) + write_jsonl( + meta_dir / "episodes.jsonl", + [ + {"episode_index": index, "length": int(length)} + for index, length in enumerate(episode_lengths) + ], + ) + write_jsonl( + meta_dir / "tasks.jsonl", + [ + {"task_index": index, "task": prompt} + for index, prompt in enumerate(task_prompts) + ], + ) + + +def _convert_legacy_split( + rows: list[dict[str, Any]], + output_dir: Path, + *, + env_name: str, + action_labels: list[str], + fps: float, + reward_field: str, + density_weighting: bool, +) -> dict[str, Any]: + output_dir.mkdir(parents=True, exist_ok=True) + prompt_to_task_index: dict[str, int] = {} + task_prompts: list[str] = [] + episode_lengths: list[int] = [] + for new_episode_idx, episode_rows in enumerate(_episode_groups(rows)): + density_weights = ( + density_weights_clean_v1(env_name, episode_rows) + if density_weighting + else [1.0] * len(episode_rows) + ) + out_rows = [] + for frame_idx, row in enumerate(episode_rows): + row["latency_raw_frames"] + prompt = str(row["prompt"]) + if prompt not in prompt_to_task_index: + prompt_to_task_index[prompt] = len(task_prompts) + task_prompts.append(prompt) + out_rows.append( + { + "image_bytes": _png_bytes(row["image"]), + "action": _legacy_row_action_vector( + row, + action_dim=len(action_labels), + ), + "timestamp": frame_idx / fps, + "episode_index": new_episode_idx, + "frame_index": frame_idx, + "task_index": prompt_to_task_index[prompt], + "done": frame_idx == len(episode_rows) - 1, + "reward": float(row[reward_field]), + "density_weight": float(density_weights[frame_idx]), + } + ) + episode_lengths.append(len(out_rows)) + episode_chunk = new_episode_idx // 1000 + _write_legacy_episode( + output_dir + / f"data/chunk-{episode_chunk:03d}/episode_{new_episode_idx:06d}.parquet", + out_rows, + action_dim=len(action_labels), + ) + _write_legacy_metadata( + output_dir, + fps=fps, + action_labels=action_labels, + episode_lengths=episode_lengths, + task_prompts=task_prompts, + ) + return { + "dataset_name": output_dir.name, + "env_name": env_name, + "episodes": len(episode_lengths), + "frames": int(sum(episode_lengths)), + "task_prompts": task_prompts, + } + + +def _convert_legacy_export( + raw_export_dir: Path, + output_dir: Path, + *, + metadata: dict[str, Any], + force: bool, +) -> dict[str, Any]: + env_name = str(metadata["env_name"]) + action_spec = action_spec_for_env(env_name) + action_labels = list(metadata["action_labels"]) + reward_field = str(metadata["reward_field"]) + rows_unit = metadata["rows_unit"] + if rows_unit != "decision_step": + raise ValueError(f"rows_unit must be 'decision_step', got {rows_unit!r}") + fps = metadata["obs_fps"] + obs_stride = metadata["obs_stride_raw_frames"] + val_output_dir = output_dir.with_name(f"{output_dir.name}__val") + if output_dir.exists() and force: + shutil.rmtree(output_dir) + if val_output_dir.exists() and force: + shutil.rmtree(val_output_dir) + train_manifest = _convert_legacy_split( + _load_rows(raw_export_dir, "train"), + output_dir, + env_name=env_name, + action_labels=action_labels, + fps=fps, + reward_field=reward_field, + density_weighting=True, + ) + val_manifest = _convert_legacy_split( + _load_rows(raw_export_dir, "val"), + val_output_dir, + env_name=env_name, + action_labels=action_labels, + fps=fps, + reward_field=reward_field, + density_weighting=False, + ) + manifest = { + **train_manifest, + "format": "starvla_lerobot_v2_image_parquet", + "source": str(raw_export_dir), + "action_layout": str(metadata["action_layout"]), + "action_labels": action_labels, + "action_dim": len(action_labels), + "active_action_dim": len(action_spec.labels), + "rows_unit": rows_unit, + "obs_fps": fps, + "obs_stride_raw_frames": obs_stride, + "validation_dataset_name": val_output_dir.name, + "validation_episodes": val_manifest["episodes"], + "validation_frames": val_manifest["frames"], + } + (output_dir / "manifest.json").write_text( + json.dumps(manifest, indent=2), + encoding="utf-8", + ) + return manifest + + +def _convert_history_game_export( + raw_export_dir: Path, + output_dir: Path, + *, + metadata: dict[str, Any], + force: bool, +) -> dict[str, Any]: + """Convert decision-v1 game rollouts with their real pre-decision state.""" + env_name = metadata["env_name"] + state_dim = _HISTORY_STATE_DIMS[env_name] + action_labels = ( + _DEADLY_MULTIBINARY_LABELS + if env_name == "deadly_corridor" else list(metadata["action_labels"]) + ) + state_labels = [f"decision_history_{index}" for index in range(state_dim)] + train_rows = _load_rows(raw_export_dir, "train") + val_rows = _load_rows(raw_export_dir, "val") + for row in [*train_rows, *val_rows]: + if len(row["state"]) != state_dim: + raise ValueError(f"{env_name} decision history has {len(row['state'])} values, expected {state_dim}") + if env_name == "deadly_corridor": + action_id = row["action_id"] + turn, move, strafe, attack = ( + action_id // 18, (action_id // 6) % 3, (action_id // 2) % 3, action_id % 2 + ) + row["action"] = [ + float(move == 1), float(move == 2), float(strafe == 1), + float(strafe == 2), float(turn == 1), float(turn == 2), + float(attack == 1), + ] + first_row = (train_rows or val_rows)[0] + with Image.open(io.BytesIO(_png_bytes(first_row["image"]))) as image: + image_shape = (image.height, image.width, 3) + val_output_dir = output_dir.with_name(f"{output_dir.name}__val") + if output_dir.exists() and force: + shutil.rmtree(output_dir) + if val_output_dir.exists() and force: + shutil.rmtree(val_output_dir) + common = { + "env_name": env_name, + "integration_name": env_name, + "action_labels": action_labels, + "active_action_dim": len(action_labels), + "state_dim": state_dim, + "state_labels": state_labels, + "context_image_count": None, + "image_shape": image_shape, + "fps": metadata["obs_fps"], + "reward_field": metadata["reward_field"], + "state_normalization": None, + "include_state": True, + "density_weighting": False, + } + train_manifest = _convert_split(train_rows, output_dir, **common) + val_manifest = _convert_split(val_rows, val_output_dir, **common) + for directory in (output_dir, val_output_dir): + modality_path = directory / "meta/modality.json" + modality = json.loads(modality_path.read_text()) + modality["annotation"] = { + "human.action.task_description": {"original_key": "task_index"} + } + modality_path.write_text(json.dumps(modality, indent=2) + "\n") + prompt_map_path, mixture_path = _write_starvla_handoff_files( + output_dir, val_output_dir, env_name=env_name, + integration_name=env_name, rows=[*train_rows, *val_rows], + ) + contract = { + "format": "starvla_lerobot_v2_decision_v1", + "source": str(raw_export_dir), + "robot_type": _ROBOT_TYPES_BY_ENV[env_name], + "action_layout": "multibinary_7" if env_name == "deadly_corridor" else metadata["action_layout"], + "source_action_layout": metadata["action_layout"], + "action_labels": action_labels, + "action_dim": len(action_labels), + "active_action_dim": len(action_labels), + "state_source": "decision_v1", + "state_dim": state_dim, + "state_labels": state_labels, + "state_normalization": {"min": [0.0] * state_dim, "max": [1.0] * state_dim}, + "action_history_decisions": metadata["action_history_decisions"], + "rows_unit": metadata["rows_unit"], + "obs_fps": metadata["obs_fps"], + "obs_stride_raw_frames": metadata["obs_stride_raw_frames"], + "latency_prompt_map_path": str(prompt_map_path), + "custom_mixtures_path": str(mixture_path), + } + if env_name in {"atlantis", "asterix"}: + # The exported student image is current RGB; four-frame stacking belongs + # to the teacher and remains recorded in the raw collection metadata. + contract.update(integration_name="gymnasium", + gymnasium_task={**metadata["gymnasium_task"], "frame_stack": 1}) + manifest = {**train_manifest, **contract, + "validation_dataset_name": val_output_dir.name, + "validation_episodes": val_manifest["episodes"], + "validation_frames": val_manifest["frames"]} + (output_dir / "manifest.json").write_text(json.dumps(manifest, indent=2) + "\n") + (val_output_dir / "manifest.json").write_text( + json.dumps({**val_manifest, **contract, "training_dataset_name": output_dir.name}, indent=2) + "\n" + ) + return manifest + + +def _write_episode( + path: Path, + rows: list[dict[str, Any]], + *, + action_dim: int, + state_dim: int, + include_state: bool, +) -> None: + """Write one LeRobot episode parquet file. + + Args: + path: Destination episode parquet path. + rows: Converted LeRobot frame rows. + action_dim: Expected action vector dimension. + state_dim: Expected state vector dimension. + include_state: Whether rows carry real state vectors. + """ + path.parent.mkdir(parents=True, exist_ok=True) + states = ( + [row["state"] for row in rows] + if include_state + else [[0.0] * state_dim for _ in rows] + ) + columns = { + "observation.image": pa.array( + [{"bytes": row["image_bytes"], "path": None} for row in rows], + type=pa.struct([("bytes", pa.binary()), ("path", pa.string())]), + ), + "observation.state": pa.array( + states, + type=pa.list_(pa.float32(), state_dim), + ), + "action": pa.array( + [row["action"] for row in rows], + type=pa.list_(pa.float32(), action_dim), + ), + "timestamp": pa.array([row["timestamp"] for row in rows], type=pa.float64()), + "episode_index": pa.array( + [row["episode_index"] for row in rows], type=pa.int64() + ), + "frame_index": pa.array([row["frame_index"] for row in rows], type=pa.int64()), + "task_index": pa.array([row["task_index"] for row in rows], type=pa.int64()), + "latency": pa.array( + [row["latency_raw_frames"] for row in rows], type=pa.int64() + ), + "latency_raw_frames": pa.array( + [row["latency_raw_frames"] for row in rows], type=pa.int64() + ), + "done": pa.array([row["done"] for row in rows], type=pa.bool_()), + "reward": pa.array([row["reward"] for row in rows], type=pa.float32()), + # Per-row loss weight for KV-memory scheme-B supervision (clean_v1 event + # balancing without deleting rows). Uniform 1.0 when density is disabled. + "density_weight": pa.array( + [row["density_weight"] for row in rows], type=pa.float32() + ), + } + if rows and "issued_command_chunk" in rows[0]: + columns["issued_command_chunk"] = pa.array( + [row["issued_command_chunk"] for row in rows], + type=pa.list_(pa.list_(pa.float32(), action_dim), len(rows[0]["issued_command_chunk"])), + ) + for mask_key in ("action_prefix_mask", "action_loss_mask"): + if rows and mask_key in rows[0]: + columns[mask_key] = pa.array( + [row[mask_key] for row in rows], + type=pa.list_(pa.bool_(), len(rows[0][mask_key])), + ) + if rows and "action_id" in rows[0]: + columns["action_id"] = pa.array( + [row["action_id"] for row in rows], type=pa.int64() + ) + if rows and "context_image_bytes" in rows[0]: + image_type = pa.struct([("bytes", pa.binary()), ("path", pa.string())]) + columns["observation.context_images"] = pa.array( + [ + [{"bytes": value, "path": None} for value in row["context_image_bytes"]] + for row in rows + ], + type=pa.list_(image_type), + ) + table = pa.table(columns) + pq.write_table(table, path) + + +def _row_action_vector( + row: dict[str, Any], + *, + active_action_dim: int, + action_dim: int, +) -> list[float]: + values = [float(value) for value in row["action"]] + return [*values, *([0.0] * (action_dim - active_action_dim))] + + +def _write_metadata( + dataset_dir: Path, + *, + fps: float, + action_labels: list[str], + episode_lengths: list[int], + task_prompts: list[str], + image_shape: tuple[int, int, int], + state_dim: int, + state_labels: list[str], + context_image_count: int | None, +) -> None: + """Write minimal LeRobot metadata files for a split dataset. + + Args: + dataset_dir: LeRobot dataset split directory. + fps: Frame rate stored in metadata. + action_labels: Canonical action labels in vector order. + episode_lengths: Number of frames per converted episode. + task_prompts: Unique prompts used as LeRobot tasks. + """ + meta_dir = dataset_dir / "meta" + meta_dir.mkdir(parents=True, exist_ok=True) + action_dim = len(action_labels) + height, width, channels = image_shape + features = { + "observation.image": { + "dtype": "image", + "shape": [height, width, channels], + "names": ["height", "width", "channel"], + "video_info": {"video.fps": fps}, + }, + "observation.state": { + "dtype": "float32", + "shape": [state_dim], + "names": state_labels, + }, + "action": {"dtype": "float32", "shape": [action_dim], "names": action_labels}, + "timestamp": {"dtype": "float64", "shape": [1]}, + "episode_index": {"dtype": "int64", "shape": [1]}, + "frame_index": {"dtype": "int64", "shape": [1]}, + "task_index": {"dtype": "int64", "shape": [1]}, + "latency": {"dtype": "int64", "shape": [1]}, + } + if context_image_count is not None: + features["observation.context_images"] = { + "dtype": "image_sequence", + "shape": [context_image_count, height, width, channels], + "names": ["time", "height", "width", "channel"], + "video_info": {"video.fps": fps}, + } + (meta_dir / "info.json").write_text( + json.dumps( + { + "codebase_version": "v2.0", + "fps": fps, + "total_episodes": len(episode_lengths), + "total_frames": int(sum(episode_lengths)), + "total_tasks": len(task_prompts), + "total_videos": 0, + "total_chunks": (len(episode_lengths) + LE_ROBOT_CHUNK_SIZE - 1) + // LE_ROBOT_CHUNK_SIZE, + "chunks_size": LE_ROBOT_CHUNK_SIZE, + "data_path": "data/chunk-{episode_chunk:03d}/episode_{episode_index:06d}.parquet", + "video_path": ( + "videos/chunk-{episode_chunk:03d}/{video_key}/" + "episode_{episode_index:06d}.mp4" + ), + "features": features, + }, + indent=2, + ), + encoding="utf-8", + ) + (meta_dir / "modality.json").write_text( + json.dumps( + { + "state": { + "game_state": { + "start": 0, + "end": state_dim, + "dtype": "float32", + "absolute": True, + "original_key": "observation.state", + } + }, + "action": { + "button": { + "start": 0, + "end": action_dim, + "dtype": "float32", + "absolute": True, + "original_key": "action", + } + }, + "video": {"image": {"original_key": "observation.image"}}, + "annotation": { + "human.task_description": {"original_key": "task_index"} + }, + }, + indent=2, + ), + encoding="utf-8", + ) + write_jsonl( + meta_dir / "episodes.jsonl", + [ + {"episode_index": idx, "length": int(length)} + for idx, length in enumerate(episode_lengths) + ], + ) + write_jsonl( + meta_dir / "tasks.jsonl", + [ + {"task_index": idx, "task": prompt} + for idx, prompt in enumerate(task_prompts) + ], + ) + + +def _convert_split( + rows: list[dict[str, Any]], + output_dir: Path, + *, + env_name: str, + integration_name: str, + action_labels: list[str], + active_action_dim: int, + state_dim: int, + state_labels: list[str], + context_image_count: int | None, + image_shape: tuple[int, int, int], + fps: float, + reward_field: str, + state_normalization: dict[str, Any] | None, + include_state: bool, + density_weighting: bool = False, +) -> dict[str, Any]: + """Convert one canonical rollout split into a LeRobot dataset directory. + + Args: + rows: Canonical rollout rows from one split. + output_dir: Destination LeRobot split directory. + env_name: Canonical benchmark environment name. + action_labels: Canonical action labels in vector order. + fps: Frame rate used for timestamps. + reward_field: Source canonical reward column. + density_weighting: When True, write per-row ``density_weight`` from clean_v1's + event balancing (train split). When False, weights are uniform 1.0 (val + split keeps an unweighted held-out loss). + + Returns: + Split manifest summary. + """ + output_dir.mkdir(parents=True, exist_ok=True) + prompt_to_task_index: dict[str, int] = {} + task_prompts: list[str] = [] + episode_lengths: list[int] = [] + for new_episode_idx, episode_rows in enumerate(_episode_groups(rows)): + density_weights = ( + density_weights_clean_v1(env_name, episode_rows) + if density_weighting + and integration_name != "gymnasium" + else [1.0] * len(episode_rows) + ) + out_rows = [] + for frame_idx, row in enumerate(episode_rows): + # Enforce the canonical rollout contract; downstream must not read the legacy ``latency`` alias. + row["latency_raw_frames"] + prompt = str(row["prompt"]) + if prompt not in prompt_to_task_index: + prompt_to_task_index[prompt] = len(task_prompts) + task_prompts.append(prompt) + out_row = { + "image_bytes": _png_bytes(row["image"]), + "action": _row_action_vector( + row, + active_action_dim=active_action_dim, + action_dim=len(action_labels), + ), + "timestamp": frame_idx / fps, + "episode_index": new_episode_idx, + "frame_index": frame_idx, + "task_index": prompt_to_task_index[prompt], + "latency_raw_frames": int(row["latency_raw_frames"]), + # ``done`` is generated only for the LeRobot target format, not for canonical rollout rows. + "done": frame_idx == len(episode_rows) - 1, + "reward": float(row[reward_field]), + "density_weight": float(density_weights[frame_idx]), + } + if include_state: + state = np.asarray(row["state"], dtype=np.float32) + if state_normalization is not None: + state_min = np.asarray(state_normalization["min"], dtype=np.float32) + state_max = np.asarray(state_normalization["max"], dtype=np.float32) + state = min_max_normalize_state(state, state_min, state_max) + out_row["state"] = state.tolist() + if "issued_command_chunk" in row: + out_row["issued_command_chunk"] = row["issued_command_chunk"] + for mask_key in ("action_prefix_mask", "action_loss_mask"): + if mask_key in row: + out_row[mask_key] = row[mask_key] + if "action_id" in row: + out_row["action_id"] = int(row["action_id"]) + if "context_images" in row and row["context_images"] is not None: + out_row["context_image_bytes"] = [ + _png_bytes(image) for image in row["context_images"] + ] + out_rows.append(out_row) + episode_lengths.append(len(out_rows)) + episode_chunk = new_episode_idx // LE_ROBOT_CHUNK_SIZE + _write_episode( + output_dir + / f"data/chunk-{episode_chunk:03d}/episode_{new_episode_idx:06d}.parquet", + out_rows, + action_dim=len(action_labels), + state_dim=state_dim, + include_state=include_state, + ) + _write_metadata( + output_dir, + fps=fps, + action_labels=action_labels, + episode_lengths=episode_lengths, + task_prompts=task_prompts, + image_shape=image_shape, + state_dim=state_dim, + state_labels=state_labels, + context_image_count=context_image_count, + ) + if rows and "issued_command_chunk" in rows[0]: + info_path = output_dir / "meta/info.json" + info = json.loads(info_path.read_text()) + info["features"]["issued_command_chunk"] = { + "dtype": "float32", "shape": [len(rows[0]["issued_command_chunk"]), len(action_labels)], + "names": action_labels, + } + for mask_key in ("action_prefix_mask", "action_loss_mask"): + if mask_key in rows[0]: + info["features"][mask_key] = { + "dtype": "bool", "shape": [len(rows[0][mask_key])], "names": None, + } + info_path.write_text(json.dumps(info, indent=2) + "\n") + return { + "dataset_name": output_dir.name, + "env_name": env_name, + "episodes": len(episode_lengths), + "frames": int(sum(episode_lengths)), + "task_prompts": task_prompts, + } + + +def convert_canonical_rollout_to_lerobot( + raw_export_dir: Path | str, + output_dir: Path | str, + *, + force: bool = False, + action_carrier: str = "native", + normalize_state: bool = True, + write_starvla_handoff: bool = True, +) -> dict[str, Any]: + """Convert a canonical rollout export into train/val LeRobot datasets. + + Args: + raw_export_dir: Directory with canonical rollout parquet files and metadata. + output_dir: Destination LeRobot train dataset directory. + force: Remove existing train/val output directories before writing. + normalize_state: Apply the existing min-max state transform when the + canonical export includes a state vector. When disabled, an + explicit ``state`` field in canonical rows is copied as-is. + write_starvla_handoff: Write the existing StarVLA prompt/mixture handoff + files and manifest fields. + + Returns: + Manifest describing the converted train dataset and validation companion. + """ + raw_export_dir = Path(raw_export_dir) + output_dir = Path(output_dir) + metadata = json.loads( + (raw_export_dir / "metadata.json").read_text(encoding="utf-8") + ) + env_name = str(metadata["env_name"]) + integration_name = ( + metadata["integration_name"] if "integration_name" in metadata else env_name + ) + if metadata.get("action_history_decisions") == 8 and env_name in _HISTORY_STATE_DIMS: + return _convert_history_game_export( + raw_export_dir, output_dir, metadata=metadata, force=force + ) + if integration_name != "gymnasium": + return _convert_legacy_export( + raw_export_dir, + output_dir, + metadata=metadata, + force=force, + ) + action_layout = metadata["action_layout"] + active_action_labels = list(metadata["action_labels"]) + if action_layout not in ( + ACTION_LAYOUT_GYMNASIUM_DISCRETE_V1, + ACTION_LAYOUT_GYMNASIUM_CONTINUOUS_V1, + ACTION_LAYOUT_G1_SONIC_ABSOLUTE_JOINT_TARGET_V1, + ): + raise ValueError(f"Unsupported Gymnasium action_layout={action_layout!r}") + gymnasium_task = metadata["gymnasium_task"] + if env_name != gymnasium_task["task_name"]: + raise ValueError( + "Canonical Gymnasium env_name does not match gymnasium_task.task_name" + ) + continuous = action_layout in ( + ACTION_LAYOUT_GYMNASIUM_CONTINUOUS_V1, + ACTION_LAYOUT_G1_SONIC_ABSOLUTE_JOINT_TARGET_V1, + ) + if continuous and action_carrier == "native": + gymnasium_task = { + **gymnasium_task, + "action_space": { + key: value + for key, value in gymnasium_task["action_space"].items() + if key not in {"openvla_carrier_dim", "openvla_padding"} + }, + } + task_action_labels = ( + gymnasium_task["action_space"]["labels"] + if continuous + else gymnasium_task["action_labels"] + ) + if active_action_labels != task_action_labels: + raise ValueError( + "Canonical Gymnasium action_labels do not match gymnasium_task action-space labels" + ) + action_labels, active_action_dim, state_dim, state_labels = _action_carrier_spec( + active_action_labels, + action_carrier, + ) + train_rows = _load_rows(raw_export_dir, "train") + val_rows = _load_rows(raw_export_dir, "val") + raw_train_rows = train_rows + raw_val_rows = val_rows + admission_filtered = ( + action_layout == ACTION_LAYOUT_G1_SONIC_ABSOLUTE_JOINT_TARGET_V1 + and any("command_admitted" in row for row in [*train_rows, *val_rows]) + ) + if admission_filtered: + train_rows = [row for row in train_rows if row["command_admitted"]] + val_rows = [row for row in val_rows if row["command_admitted"]] + uses_state = "state_space" in gymnasium_task + if not normalize_state: + uses_state = any("state" in row for row in [*train_rows, *val_rows]) + if uses_state: + if "state_space" in gymnasium_task: + state_labels = list(gymnasium_task["state_space"]["labels"]) + state_dim = len(state_labels) + else: + state_row = next(row for row in [*train_rows, *val_rows] if "state" in row) + state_dim = len(state_row["state"]) + state_labels = [f"state_{index}" for index in range(state_dim)] + reward_field = str(metadata["reward_field"]) + rows_unit = metadata["rows_unit"] + if rows_unit != "decision_step": + raise ValueError(f"rows_unit must be 'decision_step', got {rows_unit!r}") + fps = metadata["obs_fps"] + obs_stride = metadata["obs_stride_raw_frames"] + val_output_dir = output_dir.with_name(f"{output_dir.name}__val") + if output_dir.exists() and force: + shutil.rmtree(output_dir) + if val_output_dir.exists() and force: + shutil.rmtree(val_output_dir) + state_normalization = None + if uses_state and normalize_state: + train_states = np.asarray([row["state"] for row in train_rows], dtype=np.float32) + state_normalization = { + "type": "min_max", + "min": train_states.min(axis=0).tolist(), + "max": train_states.max(axis=0).tolist(), + } + first_row = (raw_train_rows or raw_val_rows)[0] + image_source = first_row["image"] + with Image.open(io.BytesIO(_png_bytes(image_source))) as image: + width, height = image.size + image_shape = (height, width, 3) + context_image_count = ( + len(first_row["context_images"]) + if "context_images" in first_row and first_row["context_images"] is not None + else None + ) + train_manifest = _convert_split( + train_rows, + output_dir, + env_name=env_name, + integration_name=integration_name, + action_labels=action_labels, + active_action_dim=active_action_dim, + state_dim=state_dim, + state_labels=state_labels, + context_image_count=context_image_count, + image_shape=image_shape, + fps=fps, + reward_field=reward_field, + state_normalization=state_normalization, + include_state=uses_state, + density_weighting=True, + ) + val_manifest = _convert_split( + val_rows, + val_output_dir, + env_name=env_name, + integration_name=integration_name, + action_labels=action_labels, + active_action_dim=active_action_dim, + state_dim=state_dim, + state_labels=state_labels, + context_image_count=context_image_count, + image_shape=image_shape, + fps=fps, + reward_field=reward_field, + state_normalization=state_normalization, + include_state=uses_state, + density_weighting=False, + ) + if "label_provenance" in metadata: + train_info = json.loads((output_dir / "meta/info.json").read_text()) + for directory in (output_dir, val_output_dir): + info_path = directory / "meta/info.json" + info = json.loads(info_path.read_text()) + info["label_provenance"] = metadata["label_provenance"] + for key in ("issued_command_chunk", "action_prefix_mask", "action_loss_mask"): + if key in train_info["features"]: + info["features"][key] = train_info["features"][key] + info_path.write_text(json.dumps(info, indent=2) + "\n") + handoff_contract = { + "format": ( + "lerobot_v2_image_parquet" + if continuous + else "starvla_lerobot_v2_image_parquet" + ), + "source": str(raw_export_dir), + "integration_name": integration_name, + "task_name": env_name, + "action_layout": str(metadata["action_layout"]), + "action_labels": active_action_labels, + "carrier_action_labels": action_labels, + "action_dim": len(action_labels), + "active_action_dim": active_action_dim, + "action_carrier": action_carrier, + "rows_unit": rows_unit, + "obs_fps": fps, + "obs_stride_raw_frames": obs_stride, + "uses_state": uses_state, + "state_dim": state_dim, + "state_labels": state_labels, + "state_normalization": state_normalization, + } + if admission_filtered: + handoff_contract.update( + { + "admission_filter": "command_admitted", + "raw_train_frames": len(raw_train_rows), + "raw_validation_frames": len(raw_val_rows), + } + ) + if write_starvla_handoff: + handoff_contract["robot_type"] = _robot_type_for_env(env_name, integration_name) + prompt_map_path, mixture_path = _write_starvla_handoff_files( + output_dir, + val_output_dir, + env_name=env_name, + integration_name=integration_name, + rows=[*train_rows, *val_rows], + ) + handoff_contract.update( + { + "latency_prompt_map_path": str(prompt_map_path), + "custom_mixtures_path": str(mixture_path), + } + ) + if gymnasium_task is not None: + handoff_contract["gymnasium_task"] = gymnasium_task + manifest = { + **train_manifest, + **handoff_contract, + "validation_dataset_name": val_output_dir.name, + "validation_episodes": val_manifest["episodes"], + "validation_frames": val_manifest["frames"], + } + validation_manifest = { + **val_manifest, + **handoff_contract, + "training_dataset_name": output_dir.name, + } + (output_dir / "manifest.json").write_text( + json.dumps(manifest, indent=2), encoding="utf-8" + ) + (val_output_dir / "manifest.json").write_text( + json.dumps(validation_manifest, indent=2), encoding="utf-8" + ) + return manifest diff --git a/latency-aware/flappy/mean_rtx3090_s42_200ep_v1/evidence/provenance/stream-raw-rgb.patch b/latency-aware/flappy/mean_rtx3090_s42_200ep_v1/evidence/provenance/stream-raw-rgb.patch new file mode 100644 index 0000000000000000000000000000000000000000..d3af4377165f253a9f090b55019c6b800d7bdc9e --- /dev/null +++ b/latency-aware/flappy/mean_rtx3090_s42_200ep_v1/evidence/provenance/stream-raw-rgb.patch @@ -0,0 +1,110 @@ +diff --git a/latency_bench/data/sf_teacher_rollout.py b/latency_bench/data/sf_teacher_rollout.py +index ef45d95b..af0a9fa9 100644 +--- a/latency_bench/data/sf_teacher_rollout.py ++++ b/latency_bench/data/sf_teacher_rollout.py +@@ -964,6 +964,11 @@ def _replay_accepted_specs( + row_buffer_size=int(row_buffer_size), + export_image_size=export_image_size, + ) ++ stream_current_images = ( ++ image_transform_config.normalized_transform() == "raw_rgb" ++ and int(context_window) <= 1 ++ and filter_preset == "none" ++ ) + image_resolution: list[int] | None = None + + def assign_specs_to_slots(slots: list[int]) -> dict[int, int]: +@@ -1051,7 +1056,10 @@ def _replay_accepted_specs( + next_raw_frame=next_decision_step * obs_stride_raw_frames, + max_steps_per_episode=int(max_steps_per_episode), + ) +- active_episode_frames[slot].append(np.array(image_frame, copy=True)) ++ if stream_current_images: ++ image_writer.write(image_frame, image_path) ++ else: ++ active_episode_frames[slot].append(np.array(image_frame, copy=True)) + row = teacher_dataset_row( + env_name=env_name, + spec=spec, +@@ -1104,7 +1112,9 @@ def _replay_accepted_specs( + ) + else: + shard_rows = [dict(row) for row in final_rows] +- if _is_demon_ghost_trail: ++ if stream_current_images: ++ image_writer.flush() ++ elif _is_demon_ghost_trail: + # Frames are pre-built ghost trail images; map directly. + ep_steps = [int(row["decision_step"]) for row in final_rows] + frames_by_decision_step = { +@@ -1118,12 +1128,13 @@ def _replay_accepted_specs( + frames=active_episode_frames[slot], + config=image_transform_config, + ) +- shard_rows = _write_filtered_episode_images( +- image_writer, +- shard_rows, +- frames_by_decision_step, +- context_window=int(context_window), +- ) ++ if not stream_current_images: ++ shard_rows = _write_filtered_episode_images( ++ image_writer, ++ shard_rows, ++ frames_by_decision_step, ++ context_window=int(context_window), ++ ) + write_episode_row_shard(output_dir, shard_rows) + progress.update(1) + replayed += 1 +diff --git a/tests/data/test_sf_teacher_dataset_export.py b/tests/data/test_sf_teacher_dataset_export.py +index e6b25233..7d5537ac 100644 +--- a/tests/data/test_sf_teacher_dataset_export.py ++++ b/tests/data/test_sf_teacher_dataset_export.py +@@ -1273,7 +1273,8 @@ def test_collect_teacher_dataset_saves_demon_attack_latest_raw_rgb_image(monkeyp + + + @pytest.mark.parametrize("export_history", [False, True]) +-def test_collect_teacher_dataset_saves_flappy_latest_raw_rgb_image(monkeypatch, tmp_path, export_history): ++@pytest.mark.parametrize("filter_preset", ["clean_v1", "none"]) ++def test_collect_teacher_dataset_saves_flappy_latest_raw_rgb_image(monkeypatch, tmp_path, export_history, filter_preset): + import torch + from gymnasium.spaces import Discrete + from PIL import Image +@@ -1307,6 +1308,9 @@ def test_collect_teacher_dataset_saves_flappy_latest_raw_rgb_image(monkeypatch, + class FakeEnv: + action_space = Discrete(2) + ++ def __init__(self, exporting): ++ self.exporting = exporting ++ + def reset(self, seed=None): + self.steps = 0 + obs = {"obs": np.zeros((1, 12, 2, 2), dtype=np.uint8)} +@@ -1323,6 +1327,9 @@ def test_collect_teacher_dataset_saves_flappy_latest_raw_rgb_image(monkeypatch, + + def step(self, action): + self.steps += 1 ++ if self.exporting and self.steps == 2 and filter_preset == "none": ++ image_path = output_dir / "images/train/episode_000000/decision_step_000000.png" ++ assert image_path.is_file(), "First frame must be on disk before the episode ends" + obs = {"obs": np.ones((1, 12, 2, 2), dtype=np.uint8)} + next_stack = np.stack([np.full((6, 5, 3), 50 + idx, dtype=np.uint8) for idx in range(4)], axis=0) + return ( +@@ -1354,7 +1361,7 @@ def test_collect_teacher_dataset_saves_flappy_latest_raw_rgb_image(monkeypatch, + exporter, + "build_env_and_policy", + lambda module, cfg, config, num_envs: { +- "envs": [FakeEnv()], ++ "envs": [FakeEnv(cfg.export_env_raw_rgb_frames)], + "env_info": object(), + "actor_critic": FakeActorCritic(), + "device": torch.device("cpu"), +@@ -1382,6 +1389,7 @@ def test_collect_teacher_dataset_saves_flappy_latest_raw_rgb_image(monkeypatch, + export_image_size=None, + image_writer_workers=0, + row_buffer_size=1, ++ filter_preset=filter_preset, + ) + train_rows = _read_shard_rows(output_dir, "train") + val_rows = _read_shard_rows(output_dir, "val") diff --git a/latency-aware/flappy/mean_rtx3090_s42_200ep_v1/evidence/provenance/vla.training-config.reference.yaml b/latency-aware/flappy/mean_rtx3090_s42_200ep_v1/evidence/provenance/vla.training-config.reference.yaml new file mode 100644 index 0000000000000000000000000000000000000000..6347967a8d8127a273bdec69b222541bd32a2296 --- /dev/null +++ b/latency-aware/flappy/mean_rtx3090_s42_200ep_v1/evidence/provenance/vla.training-config.reference.yaml @@ -0,0 +1,271 @@ +framework: + name: QwenOFT + qwenvl: + base_vlm: /data/ubuntu/lzj/models/Qwen3-VL-4B-Instruct + attn_implementation: flash_attention_2 + enable_gradient_checkpointing: true + action_model: + action_model_type: MLP + action_dim: 2 + state_dim: 24 + action_hidden_dim: 2560 + future_action_window_size: 0 + past_action_window_size: 0 + loss_type: discrete_ce + state_encoding: continuous_projector + task_objective: null + action_horizon: 1 + action_env_dim: 2 + kv_memory: + enabled: false + window: 4 + rollout_len: 8 + packed_train: false + rebased_sink: true +datasets: + vla_data: + dataset_py: lerobot_datasets + include_state: true + data_root_dir: /data/ubuntu/lzj/mean-profiling/flappy/data + data_mix: lerobot + eval_data_mix: lerobot__val + custom_mixtures_path: /data/ubuntu/lzj/mean-profiling/flappy/data/_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: 16 + 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: 8 + 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 + sample_seed: 42 + batch_size: 8 + bootstrap_repeats: 4000 +workspace_dir: WORKSPACE_DIR +run_root_dir: /data/ubuntu/lzj/mean-profiling/flappy/vla +seed: 42 +wandb_entity: dongqianyu99-zhejiang-university +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: /data/ubuntu/lzj/models/Qwen3-VL-4B-Instruct + accelerate_config: starVLA/config/deepseeds/deepspeed_zero2.yaml +launch: + use_accelerate: true + gpus: null + num_processes: 1 + dry_run: false +conda: + enabled: true + env_name: null +rl_games: + model_alias: openvla + env_eval: + image_size: 224 + frameskip: 1 + image_transform: raw_rgb + prompt_mode: raw + ghost_trail: + history_frames: 5 + gamma: 1.3 + min_alpha: 35 + scroll_px_per_step: 4.0 + ground_fraction: 0.22 + seed: 42 + fixed_episode_seeds: true + latency_seed_stride: 0 + task_seed_stride: 0 + task_description: '' + eval_parallel_envs: 5 + 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: /data/ubuntu/lzj/mean-profiling/flappy/data/lerobot/latency_prompt_map.json + 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: openvla +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-qwenoft-mean-3090-s42 +output_dir: source_archive/hist8-vla-v2-20260923/outputs/tasks/decision_history_vla_hist8_v2/train/flappy-qwenoft-hist8-scratch-s42-v2 +config_yaml: source_archive/hist8-vla-v2-20260923/outputs/tasks/decision_history_vla_hist8_v2/train/flappy-qwenoft-hist8-scratch-s42-v2/train.yaml +is_debug: false +version_id: '0.21' +wandb_name: flappy-qwenoft-mean-3090-s42 +wandb_group: mean-profiling-3090 +wandb_tags: +- flappy +- qwenoft +- hist8-v2 +- history01 +- scratch +- profile-latency +training_latency_condition: mean-profile diff --git a/latency-aware/flappy/mean_rtx3090_s42_200ep_v1/profiling/source/profile.json b/latency-aware/flappy/mean_rtx3090_s42_200ep_v1/profiling/source/profile.json new file mode 100644 index 0000000000000000000000000000000000000000..1dc1af8afd0740a8e69efc4ad41b1e3173bdf3fc --- /dev/null +++ b/latency-aware/flappy/mean_rtx3090_s42_200ep_v1/profiling/source/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": "openvla", + "n_admitted_observations": 938, + "n_capacity_drops": 1849, + "n_observation_attempts": 2787, + "per_slot_summary": { + "0": { + "admitted_count": 938, + "mean_observation_to_action_latency_ms": 75.87417450998383, + "mean_worker_service_time_ms": 75.42596991280757, + "p95_observation_to_action_latency_ms": 81.31731944084167, + "p95_worker_service_time_ms": 80.70259928703308, + "p99_worker_service_time_ms": 84.26694274425506 + } + }, + "provenance": { + "base_config": "configs/experiments/flappy/realtime/starvla_openvla_fix_latency_0_30fps_measured_warmup200_3ep.yaml", + "checkpoint_kind": "best", + "model_artifact": { + "checkpoint": "checkpoints/steps_5000_pytorch_model.pt", + "model_config": "config.full.yaml", + "path_in_repo": ".", + "repo_id": "latency-sensitive-bench/openvla_flappy_fix_latency_0", + "source": "local" + }, + "session_ids": [ + 0, + 1, + 2, + 3, + 4 + ] + }, + "sample_model_type": "hidden_regime", + "source_run_id": "20260914T122201421825Z", + "summary": { + "frame_ms": 33.333333333333336, + "max_ms": 103.8806209564209, + "mean_effective_frames": 2.2762252352995147, + "mean_ms": 75.87417450998383, + "min_ms": 69.08178091049194, + "n_samples": 938, + "p50_frames": 2.2608331131935118, + "p50_ms": 75.36110377311707, + "p90_frames": 2.404656934261322, + "p90_ms": 80.15523114204407, + "p95_frames": 2.43951958322525, + "p95_ms": 81.31731944084167, + "p99_frames": 2.557028575229644, + "p99_ms": 85.23428584098815, + "prob_latency_gt_1_frame": 1.0, + "prob_latency_gt_2_frames": 1.0, + "prob_latency_gt_3_frames": 0.0021321961620469083, + "std_ms": 3.7074819521963067 + }, + "visualization_path": "latency_profile.png", + "workload_id": "flappy" +} diff --git a/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/README.md b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/README.md new file mode 100644 index 0000000000000000000000000000000000000000..f32c6844a642e5bac904f7c3807675e79d7fd826 --- /dev/null +++ b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/README.md @@ -0,0 +1,5 @@ +# InterceptGrabFast QwenOFT H1 mean-profile raw demonstrations + +275 successful unique-seed episodes,250train/25val,11119/1092rows. EEFstate7/nativecontinuousaction7,two128x128RGB cameras. Fixedmean99.05021289731565ms from archivedRTX3090 QwenOFT profile,20Hz two raw frames (100ms effective),H1/noDAgger. Original native collection/NPZ->RLDS->LeRobot conversion complete and full numeric/application latency/seed audit passed. + +RawNPZ arrays are published under demonstrations/raw/raw_npz. The registered dataset config presents episode manifests pointing to these files; it does not expose per-step images/actions through the viewer. Source converters/expected metadata/authoritativeTFDSsplit evidence are included, converted training data remain on host. Teacher full2441iterations/149975040actualsteps under original150Mconfigured floorbudget completed and inferencecheckpoint was safely published/readback verified. Data preparation is not completed VLA training or independent benchmark acceptance. diff --git a/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/conversion/README.md b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/conversion/README.md new file mode 100644 index 0000000000000000000000000000000000000000..4843c3bdc01b69a601cc4a03184cce610acc4256 --- /dev/null +++ b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/conversion/README.md @@ -0,0 +1,3 @@ +Raw source episodes retain the original NPZ format and two128x128RGB cameras (6channels), EEFproprio7 and issuedaction7, success/seed/latency/application evidence. ManifestNPZ paths are relative to demonstrations/raw. The dataset card registers per-episode metadata manifests; rawNPZ arrays are separate payload files, not viewer rows. + +Reproduce using the original MIKASA NPZ->RLDS->LeRobot converters at the recorded revisions and supplied source. Build all275episodes into RLDS and convert train[:250] and train[250:275] with fps20, shuffle_files=false. dataset.split.audit.json records authoritative originalTFDS episode_metadata.file_path mapping; do not split NPZnumeric/lex order instead. Converted LeRobot data are retained on the host, expected metadata/taskcontracts included here. diff --git a/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/conversion/expected/lerobot/manifest.json b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/conversion/expected/lerobot/manifest.json new file mode 100644 index 0000000000000000000000000000000000000000..fb1e206169bd83b42b598fe652aefa38ff69e81b --- /dev/null +++ b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/conversion/expected/lerobot/manifest.json @@ -0,0 +1,10 @@ +{ + "format": "latency_bench_starvla_h1_v1", + "task": "mikasa_intercept_grab_fast", + "robot_type": "latency_mikasa_h1", + "state_dim": 7, + "action_dim": 7, + "action_horizon": 1, + "vla_fps": 20.0, + "dagger": false +} diff --git a/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/conversion/expected/lerobot/meta/episodes/chunk-000/file-000.parquet b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/conversion/expected/lerobot/meta/episodes/chunk-000/file-000.parquet new file mode 100644 index 0000000000000000000000000000000000000000..53b74293ea024e29c6e940f729b478fefe50ffbb --- /dev/null +++ b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/conversion/expected/lerobot/meta/episodes/chunk-000/file-000.parquet @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:5d0ecb7b2409699ddd06a99ecd4863c6fc20cfdd4bef55473c68baf35aa86a94 +size 1070540 diff --git a/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/conversion/expected/lerobot/meta/info.json b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/conversion/expected/lerobot/meta/info.json new file mode 100644 index 0000000000000000000000000000000000000000..abbac5361e17f2affa073d75b688f314a8268f68 --- /dev/null +++ b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/conversion/expected/lerobot/meta/info.json @@ -0,0 +1,111 @@ +{ + "codebase_version": "v3.0", + "robot_type": "mikasa_robo", + "total_episodes": 250, + "total_frames": 11119, + "total_tasks": 1, + "chunks_size": 1000, + "data_files_size_in_mb": 100, + "video_files_size_in_mb": 200, + "fps": 20, + "splits": { + "train": "0:250" + }, + "data_path": "data/chunk-{chunk_index:03d}/file-{file_index:03d}.parquet", + "video_path": "videos/{video_key}/chunk-{chunk_index:03d}/file-{file_index:03d}.mp4", + "features": { + "observation.images.top": { + "dtype": "video", + "shape": [ + 128, + 128, + 3 + ], + "names": [ + "height", + "width", + "channel" + ], + "info": { + "video.height": 128, + "video.width": 128, + "video.codec": "av1", + "video.pix_fmt": "yuv420p", + "video.is_depth_map": false, + "video.fps": 20, + "video.channels": 3, + "has_audio": false + } + }, + "observation.state": { + "dtype": "float32", + "shape": [ + 7 + ] + }, + "action": { + "dtype": "float32", + "shape": [ + 7 + ] + }, + "observation.images.wrist": { + "dtype": "video", + "shape": [ + 128, + 128, + 3 + ], + "names": [ + "height", + "width", + "channel" + ], + "info": { + "video.height": 128, + "video.width": 128, + "video.codec": "av1", + "video.pix_fmt": "yuv420p", + "video.is_depth_map": false, + "video.fps": 20, + "video.channels": 3, + "has_audio": false + } + }, + "timestamp": { + "dtype": "float32", + "shape": [ + 1 + ], + "names": null + }, + "frame_index": { + "dtype": "int64", + "shape": [ + 1 + ], + "names": null + }, + "episode_index": { + "dtype": "int64", + "shape": [ + 1 + ], + "names": null + }, + "index": { + "dtype": "int64", + "shape": [ + 1 + ], + "names": null + }, + "task_index": { + "dtype": "int64", + "shape": [ + 1 + ], + "names": null + } + } +} \ No newline at end of file diff --git a/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/conversion/expected/lerobot/meta/modality.json b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/conversion/expected/lerobot/meta/modality.json new file mode 100644 index 0000000000000000000000000000000000000000..0f95ae7bf021c31b744ff93206cdac99e616c0db --- /dev/null +++ b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/conversion/expected/lerobot/meta/modality.json @@ -0,0 +1,33 @@ +{ + "state": { + "proprio": { + "start": 0, + "end": 7, + "dtype": "float32", + "absolute": true, + "original_key": "observation.state" + } + }, + "action": { + "target": { + "start": 0, + "end": 7, + "dtype": "float32", + "absolute": false, + "original_key": "action" + } + }, + "video": { + "top": { + "original_key": "observation.images.top" + }, + "wrist": { + "original_key": "observation.images.wrist" + } + }, + "annotation": { + "human.task_description": { + "original_key": "task_index" + } + } +} diff --git a/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/conversion/expected/lerobot/meta/stats.json b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/conversion/expected/lerobot/meta/stats.json new file mode 100644 index 0000000000000000000000000000000000000000..c1fef365c29f3c23192a36b0e4e2d0a74b738a61 --- /dev/null +++ b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/conversion/expected/lerobot/meta/stats.json @@ -0,0 +1,650 @@ +{ + "index": { + "min": [ + 0 + ], + "max": [ + 11118 + ], + "mean": [ + 5559.0 + ], + "std": [ + 3209.77880857856 + ], + "count": [ + 11119 + ], + "q01": [ + 5537.131936325097 + ], + "q10": [ + 5541.1003881787055 + ], + "q50": [ + 5558.968251029768 + ], + "q90": [ + 5576.890709941632 + ], + "q99": [ + 5580.8641269705 + ] + }, + "observation.images.top": { + "min": [ + [ + [ + 0.03529411764705882 + ] + ], + [ + [ + 0.0 + ] + ], + [ + [ + 0.0 + ] + ] + ], + "max": [ + [ + [ + 1.0 + ] + ], + [ + [ + 1.0 + ] + ], + [ + [ + 1.0 + ] + ] + ], + "mean": [ + [ + [ + 0.6485406525989132 + ] + ], + [ + [ + 0.4533750770624158 + ] + ], + [ + [ + 0.35021969313178936 + ] + ] + ], + "std": [ + [ + [ + 0.001838113351052855 + ] + ], + [ + [ + 0.0011173426445971947 + ] + ], + [ + [ + 0.0023120831387856765 + ] + ] + ], + "count": [ + 11119 + ], + "q01": [ + [ + [ + 0.18479078305335367 + ] + ], + [ + [ + 0.18346944180510175 + ] + ], + [ + [ + 0.16332735674425988 + ] + ] + ], + "q10": [ + [ + [ + 0.4196659760630178 + ] + ], + [ + [ + 0.3401429863728414 + ] + ], + [ + [ + 0.21167452884299942 + ] + ] + ], + "q50": [ + [ + [ + 0.6594093580201142 + ] + ], + [ + [ + 0.4392742414188981 + ] + ], + [ + [ + 0.31061132008041464 + ] + ] + ], + "q90": [ + [ + [ + 0.8902053074224021 + ] + ], + [ + [ + 0.5684512257090099 + ] + ], + [ + [ + 0.47447001955547163 + ] + ] + ], + "q99": [ + [ + [ + 0.9999339309571906 + ] + ], + [ + [ + 0.7886652034059227 + ] + ], + [ + [ + 0.7840322508679234 + ] + ] + ] + }, + "frame_index": { + "min": [ + 0 + ], + "max": [ + 54 + ], + "mean": [ + 21.868063674790907 + ], + "std": [ + 13.061999586794009 + ], + "count": [ + 11119 + ], + "q01": [ + -1.0000000000000011e-10 + ], + "q10": [ + 3.96845185350212 + ], + "q50": [ + 21.8282617141828 + ], + "q90": [ + 39.758773616414324 + ], + "q99": [ + 43.73219064530195 + ] + }, + "action": { + "min": [ + -0.6572320461273193, + -1.0, + -0.8218780755996704, + -0.7686842083930969, + -0.9920897483825684, + -0.30364304780960083, + -1.0 + ], + "max": [ + 0.30179348587989807, + 0.7795099020004272, + 0.2688712179660797, + 0.8059309720993042, + 0.5333064198493958, + 1.0, + 1.0 + ], + "mean": [ + -0.0486972652612968, + 0.2488648415078191, + -0.1647999347742486, + -0.03260910847532797, + -0.5016916031105842, + 0.44703683922890974, + 0.679378104275299 + ], + "std": [ + 0.1079770799196133, + 0.4720366331096493, + 0.14475310850547873, + 0.2849435398821666, + 0.3181208643331281, + 0.35324051366204945, + 0.634955737492212 + ], + "count": [ + 11119 + ], + "q01": [ + -0.3474905894120581, + -0.9837311134405708, + -0.7067701809804469, + -0.4551988062331955, + -0.8763280069670432, + -0.09687308409449655, + -0.8752664054498279 + ], + "q10": [ + -0.18452646399416214, + -0.5284071915049338, + -0.2615673164263056, + -0.3223622702715456, + -0.8172078522568358, + -0.021898940225810338, + -0.5464212795469396 + ], + "q50": [ + -0.019935300338342253, + 0.5034426050908124, + -0.19039602586933366, + -0.11193197938282827, + -0.6001039694507877, + 0.40690699979838474, + 0.9997492765457423 + ], + "q90": [ + 0.03249546505181804, + 0.6548773488284527, + 0.008307202229644338, + 0.429210709174371, + 0.03542838243070572, + 0.9156851816728028, + 0.9999498553091489 + ], + "q99": [ + 0.07052272752156326, + 0.6879038016183283, + 0.21105365860368472, + 0.6931296005086279, + 0.1708760927429502, + 0.9580082332325738, + 0.9999949855309147 + ] + }, + "episode_index": { + "min": [ + 0 + ], + "max": [ + 249 + ], + "mean": [ + 124.63413976076986 + ], + "std": [ + 72.17226003330649 + ], + "count": [ + 11119 + ], + "q01": [ + 124.63413976076986 + ], + "q10": [ + 124.63413976076986 + ], + "q50": [ + 124.63413976076986 + ], + "q90": [ + 124.63413976076986 + ], + "q99": [ + 124.63413976076986 + ] + }, + "timestamp": { + "min": [ + 0.0 + ], + "max": [ + 2.7 + ], + "mean": [ + 1.0934031837395415 + ], + "std": [ + 0.653099979339701 + ], + "count": [ + 11119 + ], + "q01": [ + -1.0000000000000011e-10 + ], + "q10": [ + 0.19842259259736267 + ], + "q50": [ + 1.0918149635757237 + ], + "q90": [ + 1.9879386808984185 + ], + "q99": [ + 2.186609532360077 + ] + }, + "task_index": { + "min": [ + 0 + ], + "max": [ + 0 + ], + "mean": [ + 0.0 + ], + "std": [ + 0.0 + ], + "count": [ + 11119 + ], + "q01": [ + 3.9999999999994236e-16 + ], + "q10": [ + 3.999999999999418e-15 + ], + "q50": [ + 1.9999999999997103e-14 + ], + "q90": [ + 3.599999999999476e-14 + ], + "q99": [ + 3.9599999999994254e-14 + ] + }, + "observation.images.wrist": { + "min": [ + [ + [ + 0.0 + ] + ], + [ + [ + 0.0 + ] + ], + [ + [ + 0.0 + ] + ] + ], + "max": [ + [ + [ + 1.0 + ] + ], + [ + [ + 1.0 + ] + ], + [ + [ + 1.0 + ] + ] + ], + "mean": [ + [ + [ + 0.6753520986140742 + ] + ], + [ + [ + 0.4108834469111261 + ] + ], + [ + [ + 0.27539044824561293 + ] + ] + ], + "std": [ + [ + [ + 0.005396488206793373 + ] + ], + [ + [ + 0.006096790546071799 + ] + ], + [ + [ + 0.007077412543809987 + ] + ] + ], + "count": [ + 11119 + ], + "q01": [ + [ + [ + 0.06980259147576204 + ] + ], + [ + [ + 0.0005697344669048271 + ] + ], + [ + [ + 0.0005661974275424577 + ] + ] + ], + "q10": [ + [ + [ + 0.4975025611125965 + ] + ], + [ + [ + 0.2866040062892914 + ] + ], + [ + [ + 0.18232198144855516 + ] + ] + ], + "q50": [ + [ + [ + 0.6797379431508024 + ] + ], + [ + [ + 0.4138611681496012 + ] + ], + [ + [ + 0.2577839862160154 + ] + ] + ], + "q90": [ + [ + [ + 0.887266384246872 + ] + ], + [ + [ + 0.5520007460819158 + ] + ], + [ + [ + 0.36778023892440403 + ] + ] + ], + "q99": [ + [ + [ + 0.9998971667357421 + ] + ], + [ + [ + 0.8483496410002673 + ] + ], + [ + [ + 0.8513326622435947 + ] + ] + ] + }, + "observation.state": { + "min": [ + -0.37446925044059753, + -0.29720592498779297, + 0.016940969973802567, + -3.140481948852539, + -0.10707131028175354, + -1.2001688480377197, + 0.03238735347986221 + ], + "max": [ + 0.014208495616912842, + 0.5909395217895508, + 0.3558541536331177, + 3.1415467262268066, + 0.9420600533485413, + 0.09770228713750839, + 0.07999999821186066 + ], + "mean": [ + -0.14554798037731076, + 0.07433982721218718, + 0.14363656104433833, + 2.7008670382881794, + 0.45680970890571965, + -0.482290827879021, + 0.06972623075571235 + ], + "std": [ + 0.0767860073989027, + 0.2198829893890316, + 0.09388110819094506, + 1.109086251929196, + 0.30128461491380076, + 0.30068838668820946, + 0.016448576528778097 + ], + "count": [ + 11119 + ], + "q01": [ + -0.19463537064323067, + -0.1983900279348108, + 0.019201690046151382, + -0.13862184035562544, + -0.004217044630985364, + -0.8413448800890764, + 0.03574505091894174 + ], + "q10": [ + -0.17744038977297596, + -0.1817026086165002, + 0.02450567153784602, + 2.745659745859899, + 0.02050611639289151, + -0.8118146904129829, + 0.037329416763955785 + ], + "q50": [ + -0.1530566211290305, + 0.016341574399163283, + 0.13827744013411988, + 2.899938738553018, + 0.48790259276944664, + -0.5451134645766802, + 0.07989394938486216 + ], + "q90": [ + -0.08679016741035048, + 0.4258482108855702, + 0.27581075693484297, + 3.066509811781068, + 0.8140479788625072, + -0.0010797752735122097, + 0.07999797948711193 + ], + "q99": [ + -0.06210018178801098, + 0.43615247348052294, + 0.31855592609088873, + 3.128159362672204, + 0.8309532630578667, + 0.01833643887417565, + 0.07999979633938568 + ] + } +} \ No newline at end of file diff --git a/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/conversion/expected/lerobot/meta/stats_gr00t.json b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/conversion/expected/lerobot/meta/stats_gr00t.json new file mode 100644 index 0000000000000000000000000000000000000000..4d02a57cf8410bb80060135f9a0962d9d317aa5c --- /dev/null +++ b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/conversion/expected/lerobot/meta/stats_gr00t.json @@ -0,0 +1,120 @@ +{ + "__format_version": 2, + "__cache_config": { + "mode": "abs" + }, + "statistics": { + "observation.state": { + "mean": [ + -0.14554812014102936, + 0.07433967292308807, + 0.1436367779970169, + 2.700871706008911, + 0.45681115984916687, + -0.48229172825813293, + 0.06972828507423401 + ], + "std": [ + 0.07678605616092682, + 0.21988309919834137, + 0.09388109296560287, + 1.109084963798523, + 0.3012843728065491, + 0.3006882667541504, + 0.0164483655244112 + ], + "min": [ + -0.37446925044059753, + -0.29720592498779297, + 0.016940969973802567, + -3.140481948852539, + -0.10707131028175354, + -1.2001688480377197, + 0.03238735347986221 + ], + "max": [ + 0.014208495616912842, + 0.5909395217895508, + 0.3558541536331177, + 3.1415467262268066, + 0.9420600533485413, + 0.09770228713750839, + 0.07999999821186066 + ], + "q01": [ + -0.3139808464050293, + -0.25229972660541533, + 0.0188405978679657, + -3.134360270500183, + -0.04521620273590088, + -1.0301372480392457, + 0.03585549935698509 + ], + "q99": [ + -0.013631248474121185, + 0.5275326526165007, + 0.33348673582077026, + 3.1367013454437256, + 0.8922865676879882, + 0.056985778063535636, + 0.07999999821186066 + ] + }, + "action": { + "mean": [ + -0.048697348684072495, + 0.24886493384838104, + -0.16479986906051636, + -0.032609131187200546, + -0.501691997051239, + 0.44703611731529236, + 0.6793776154518127 + ], + "std": [ + 0.1079770028591156, + 0.47203660011291504, + 0.14475314319133759, + 0.2849433720111847, + 0.318120539188385, + 0.35324087738990784, + 0.6349603533744812 + ], + "min": [ + -0.6572320461273193, + -1.0, + -0.8218780755996704, + -0.7686842083930969, + -0.9920897483825684, + -0.30364304780960083, + -1.0 + ], + "max": [ + 0.30179348587989807, + 0.7795099020004272, + 0.2688712179660797, + 0.8059309720993042, + 0.5333064198493958, + 1.0, + 1.0 + ], + "q01": [ + -0.4552440994977951, + -1.0, + -0.7419710528850555, + -0.475678391456604, + -0.9305388247966766, + -0.20442348539829253, + -0.8981475746631622 + ], + "q99": [ + 0.1125871029496192, + 0.7402709639072418, + 0.22590837419033025, + 0.7005072689056396, + 0.16810789704322815, + 1.0, + 1.0 + ] + } + } +} diff --git a/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/conversion/expected/lerobot/meta/tasks.parquet b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/conversion/expected/lerobot/meta/tasks.parquet new file mode 100644 index 0000000000000000000000000000000000000000..9c55ec41283f569952a3f11813c4924db9e53c7b --- /dev/null +++ b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/conversion/expected/lerobot/meta/tasks.parquet @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:29b49531995b967e7b4afc52add5605c15e84ad1dd445ca06dccff69df98e5f5 +size 2351 diff --git a/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/conversion/expected/lerobot/task_contract.json b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/conversion/expected/lerobot/task_contract.json new file mode 100644 index 0000000000000000000000000000000000000000..76ac5a2d1a2e11769474504874b6e909be02e9b8 --- /dev/null +++ b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/conversion/expected/lerobot/task_contract.json @@ -0,0 +1,131 @@ +{ + "format": "latency_bench_starvla_h1_v1", + "task": "mikasa_intercept_grab_fast", + "robot_type": "latency_mikasa_h1", + "state_dim": 7, + "action_dim": 7, + "action_horizon": 1, + "fps": 20.0, + "state_key": "observation.state", + "action_key": "action", + "video_keys": [ + "video.top", + "video.wrist" + ], + "auxiliary": {}, + "normalization": { + "state": { + "mean": [ + -0.14554812014102936, + 0.07433967292308807, + 0.1436367779970169, + 2.700871706008911, + 0.45681115984916687, + -0.48229172825813293, + 0.06972828507423401 + ], + "std": [ + 0.07678605616092682, + 0.21988309919834137, + 0.09388109296560287, + 1.109084963798523, + 0.3012843728065491, + 0.3006882667541504, + 0.0164483655244112 + ], + "min": [ + -0.37446925044059753, + -0.29720592498779297, + 0.016940969973802567, + -3.140481948852539, + -0.10707131028175354, + -1.2001688480377197, + 0.03238735347986221 + ], + "max": [ + 0.014208495616912842, + 0.5909395217895508, + 0.3558541536331177, + 3.1415467262268066, + 0.9420600533485413, + 0.09770228713750839, + 0.07999999821186066 + ], + "q01": [ + -0.3139808464050293, + -0.25229972660541533, + 0.0188405978679657, + -3.134360270500183, + -0.04521620273590088, + -1.0301372480392457, + 0.03585549935698509 + ], + "q99": [ + -0.013631248474121185, + 0.5275326526165007, + 0.33348673582077026, + 3.1367013454437256, + 0.8922865676879882, + 0.056985778063535636, + 0.07999999821186066 + ] + }, + "action": { + "mean": [ + -0.048697348684072495, + 0.24886493384838104, + -0.16479986906051636, + -0.032609131187200546, + -0.501691997051239, + 0.44703611731529236, + 0.6793776154518127 + ], + "std": [ + 0.1079770028591156, + 0.47203660011291504, + 0.14475314319133759, + 0.2849433720111847, + 0.318120539188385, + 0.35324087738990784, + 0.6349603533744812 + ], + "min": [ + -0.6572320461273193, + -1.0, + -0.8218780755996704, + -0.7686842083930969, + -0.9920897483825684, + -0.30364304780960083, + -1.0 + ], + "max": [ + 0.30179348587989807, + 0.7795099020004272, + 0.2688712179660797, + 0.8059309720993042, + 0.5333064198493958, + 1.0, + 1.0 + ], + "q01": [ + -0.4552440994977951, + -1.0, + -0.7419710528850555, + -0.475678391456604, + -0.9305388247966766, + -0.20442348539829253, + -0.8981475746631622 + ], + "q99": [ + 0.1125871029496192, + 0.7402709639072418, + 0.22590837419033025, + 0.7005072689056396, + 0.16810789704322815, + 1.0, + 1.0 + ] + } + }, + "prompt": "Intercept the rolling ball and grasp it to stop it." +} diff --git a/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/conversion/expected/lerobot__val/manifest.json b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/conversion/expected/lerobot__val/manifest.json new file mode 100644 index 0000000000000000000000000000000000000000..fb1e206169bd83b42b598fe652aefa38ff69e81b --- /dev/null +++ b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/conversion/expected/lerobot__val/manifest.json @@ -0,0 +1,10 @@ +{ + "format": "latency_bench_starvla_h1_v1", + "task": "mikasa_intercept_grab_fast", + "robot_type": "latency_mikasa_h1", + "state_dim": 7, + "action_dim": 7, + "action_horizon": 1, + "vla_fps": 20.0, + "dagger": false +} diff --git a/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/conversion/expected/lerobot__val/meta/episodes/chunk-000/file-000.parquet b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/conversion/expected/lerobot__val/meta/episodes/chunk-000/file-000.parquet new file mode 100644 index 0000000000000000000000000000000000000000..27e7b92caf764614b9e6db5feab0529a9ca216a3 --- /dev/null +++ b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/conversion/expected/lerobot__val/meta/episodes/chunk-000/file-000.parquet @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:725ec4029a221dc57f839cfbca5698355ccfdf91f9fdea11d6db570017e89a76 +size 143349 diff --git a/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/conversion/expected/lerobot__val/meta/info.json b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/conversion/expected/lerobot__val/meta/info.json new file mode 100644 index 0000000000000000000000000000000000000000..c3fa97c1c48aa1e489f3d259b9e01a2b1202a77a --- /dev/null +++ b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/conversion/expected/lerobot__val/meta/info.json @@ -0,0 +1,111 @@ +{ + "codebase_version": "v3.0", + "robot_type": "mikasa_robo", + "total_episodes": 25, + "total_frames": 1092, + "total_tasks": 1, + "chunks_size": 1000, + "data_files_size_in_mb": 100, + "video_files_size_in_mb": 200, + "fps": 20, + "splits": { + "train": "0:25" + }, + "data_path": "data/chunk-{chunk_index:03d}/file-{file_index:03d}.parquet", + "video_path": "videos/{video_key}/chunk-{chunk_index:03d}/file-{file_index:03d}.mp4", + "features": { + "observation.images.top": { + "dtype": "video", + "shape": [ + 128, + 128, + 3 + ], + "names": [ + "height", + "width", + "channel" + ], + "info": { + "video.height": 128, + "video.width": 128, + "video.codec": "av1", + "video.pix_fmt": "yuv420p", + "video.is_depth_map": false, + "video.fps": 20, + "video.channels": 3, + "has_audio": false + } + }, + "observation.state": { + "dtype": "float32", + "shape": [ + 7 + ] + }, + "action": { + "dtype": "float32", + "shape": [ + 7 + ] + }, + "observation.images.wrist": { + "dtype": "video", + "shape": [ + 128, + 128, + 3 + ], + "names": [ + "height", + "width", + "channel" + ], + "info": { + "video.height": 128, + "video.width": 128, + "video.codec": "av1", + "video.pix_fmt": "yuv420p", + "video.is_depth_map": false, + "video.fps": 20, + "video.channels": 3, + "has_audio": false + } + }, + "timestamp": { + "dtype": "float32", + "shape": [ + 1 + ], + "names": null + }, + "frame_index": { + "dtype": "int64", + "shape": [ + 1 + ], + "names": null + }, + "episode_index": { + "dtype": "int64", + "shape": [ + 1 + ], + "names": null + }, + "index": { + "dtype": "int64", + "shape": [ + 1 + ], + "names": null + }, + "task_index": { + "dtype": "int64", + "shape": [ + 1 + ], + "names": null + } + } +} \ No newline at end of file diff --git a/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/conversion/expected/lerobot__val/meta/modality.json b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/conversion/expected/lerobot__val/meta/modality.json new file mode 100644 index 0000000000000000000000000000000000000000..0f95ae7bf021c31b744ff93206cdac99e616c0db --- /dev/null +++ b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/conversion/expected/lerobot__val/meta/modality.json @@ -0,0 +1,33 @@ +{ + "state": { + "proprio": { + "start": 0, + "end": 7, + "dtype": "float32", + "absolute": true, + "original_key": "observation.state" + } + }, + "action": { + "target": { + "start": 0, + "end": 7, + "dtype": "float32", + "absolute": false, + "original_key": "action" + } + }, + "video": { + "top": { + "original_key": "observation.images.top" + }, + "wrist": { + "original_key": "observation.images.wrist" + } + }, + "annotation": { + "human.task_description": { + "original_key": "task_index" + } + } +} diff --git a/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/conversion/expected/lerobot__val/meta/stats.json b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/conversion/expected/lerobot__val/meta/stats.json new file mode 100644 index 0000000000000000000000000000000000000000..ba3dd27ce4bfb74226a662ddd3b69871ba36ec77 --- /dev/null +++ b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/conversion/expected/lerobot__val/meta/stats.json @@ -0,0 +1,650 @@ +{ + "index": { + "min": [ + 0 + ], + "max": [ + 1091 + ], + "mean": [ + 545.5 + ], + "std": [ + 315.2331148002486 + ], + "count": [ + 1092 + ], + "q01": [ + 524.0045787544786 + ], + "q10": [ + 527.9594761537644 + ], + "q50": [ + 545.4416144688643 + ], + "q90": [ + 563.0405238462354 + ], + "q99": [ + 566.9916115788548 + ] + }, + "action": { + "min": [ + -0.6437053680419922, + -1.0, + -0.8190652132034302, + -0.7167630195617676, + -0.9770606160163879, + -0.22268825769424438, + -0.9664353728294373 + ], + "max": [ + 0.24019211530685425, + 0.7757453322410583, + 0.2596358060836792, + 0.7934277653694153, + 0.4550458490848541, + 1.0, + 1.0 + ], + "mean": [ + -0.044902275770484945, + 0.25167198022042875, + -0.1672952418444139, + -0.03629667643498588, + -0.5173969276167534, + 0.47611199417612055, + 0.6867632562970067 + ], + "std": [ + 0.10045907401412776, + 0.4802984490739245, + 0.14755421993037668, + 0.28819737142657503, + 0.31690076092325226, + 0.33510449217677185, + 0.6320979889773323 + ], + "count": [ + 1092 + ], + "q01": [ + -0.3226347790532933, + -0.9890358958379689, + -0.7189640540888894, + -0.43125362347661356, + -0.8930576148923937, + -0.07413713001192382, + -0.8722216905036688 + ], + "q10": [ + -0.16914160024210978, + -0.5395369176735808, + -0.2675455926484241, + -0.3215394225599238, + -0.827590303734327, + -0.011211257385970135, + -0.5504581930163579 + ], + "q50": [ + -0.018317628112850173, + 0.5145627577335407, + -0.1920086235190049, + -0.11589103520953611, + -0.6132637387612365, + 0.43323295196388667, + 0.9997531316191769 + ], + "q90": [ + 0.030610135687521967, + 0.6642192684195853, + 0.005590429257997226, + 0.4363410734868312, + 0.014816537979758264, + 0.9402452187802328, + 0.9999506263238355 + ], + "q99": [ + 0.07559209882827551, + 0.6949022775375362, + 0.21196211619978095, + 0.6954575046378396, + 0.16148332992012548, + 0.9816847987695525, + 0.9999950626323834 + ] + }, + "task_index": { + "min": [ + 0 + ], + "max": [ + 0 + ], + "mean": [ + 0.0 + ], + "std": [ + 0.0 + ], + "count": [ + 1092 + ], + "q01": [ + 3.9999999999994176e-16 + ], + "q10": [ + 3.999999999999417e-15 + ], + "q50": [ + 1.9999999999997088e-14 + ], + "q90": [ + 3.5999999999994766e-14 + ], + "q99": [ + 3.959999999999424e-14 + ] + }, + "episode_index": { + "min": [ + 0 + ], + "max": [ + 24 + ], + "mean": [ + 11.853479853479854 + ], + "std": [ + 7.147104025353752 + ], + "count": [ + 1092 + ], + "q01": [ + 11.853479853479854 + ], + "q10": [ + 11.853479853479858 + ], + "q50": [ + 11.853479853479874 + ], + "q90": [ + 11.853479853479891 + ], + "q99": [ + 11.853479853479895 + ] + }, + "frame_index": { + "min": [ + 0 + ], + "max": [ + 50 + ], + "mean": [ + 21.495421245421245 + ], + "std": [ + 12.87346765682025 + ], + "count": [ + 1092 + ], + "q01": [ + -1.0000000000000002e-10 + ], + "q10": [ + 3.954897399185772 + ], + "q50": [ + 21.42987307692279 + ], + "q90": [ + 39.03594509165672 + ], + "q99": [ + 42.98703282427581 + ] + }, + "observation.images.top": { + "min": [ + [ + [ + 0.03529411764705882 + ] + ], + [ + [ + 0.0 + ] + ], + [ + [ + 0.0 + ] + ] + ], + "max": [ + [ + [ + 1.0 + ] + ], + [ + [ + 1.0 + ] + ], + [ + [ + 1.0 + ] + ] + ], + "mean": [ + [ + [ + 0.6484179212299902 + ] + ], + [ + [ + 0.45346822181772345 + ] + ], + [ + [ + 0.35043059967376367 + ] + ] + ], + "std": [ + [ + [ + 0.001678829301300219 + ] + ], + [ + [ + 0.0009520436874405448 + ] + ], + [ + [ + 0.001865714394213745 + ] + ] + ], + "count": [ + 1092 + ], + "q01": [ + [ + [ + 0.18452921788009186 + ] + ], + [ + [ + 0.18357941364205552 + ] + ], + [ + [ + 0.16316797668666713 + ] + ] + ], + "q10": [ + [ + [ + 0.41966762555891324 + ] + ], + [ + [ + 0.33997207749429537 + ] + ], + [ + [ + 0.21168801479309968 + ] + ] + ], + "q50": [ + [ + [ + 0.6596620717524874 + ] + ], + [ + [ + 0.4392744982337595 + ] + ], + [ + [ + 0.3107257932405243 + ] + ] + ], + "q90": [ + [ + [ + 0.8902327522591846 + ] + ], + [ + [ + 0.5690293915114858 + ] + ], + [ + [ + 0.4744742636765667 + ] + ] + ], + "q99": [ + [ + [ + 0.9999332799570102 + ] + ], + [ + [ + 0.7903927949779378 + ] + ], + [ + [ + 0.7869821613580817 + ] + ] + ] + }, + "observation.state": { + "min": [ + -0.36666107177734375, + -0.28134918212890625, + 0.017842963337898254, + -3.141038656234741, + -0.06073141098022461, + -1.0314348936080933, + 0.03428912162780762 + ], + "max": [ + -0.03523063659667969, + 0.5806169509887695, + 0.3414648771286011, + 3.1413722038269043, + 0.9162643551826477, + 0.08547469973564148, + 0.07999999821186066 + ], + "mean": [ + -0.13605183630076412, + 0.06997808432548723, + 0.1438083655231602, + 2.6396909087568847, + 0.4466764385446961, + -0.492265733279588, + 0.06992356316101202 + ], + "std": [ + 0.06379907982945246, + 0.21989607756916, + 0.09346858016015534, + 1.2415622460027917, + 0.3019570426186774, + 0.310104879025474, + 0.016306935241816624 + ], + "count": [ + 1092 + ], + "q01": [ + -0.17616428863325398, + -0.2002109477548918, + 0.01926403638016392, + -0.652767865649073, + -0.009352170785393518, + -0.8726087084565407, + 0.035856180189811925 + ], + "q10": [ + -0.16185229060686987, + -0.18317582884278058, + 0.024263060980070472, + 2.1361652505921795, + 0.014411605406687301, + -0.8412944837449453, + 0.03713223856426911 + ], + "q50": [ + -0.1430205789699432, + 0.011594377744656343, + 0.13905346139433075, + 2.89240241028887, + 0.46928471167172703, + -0.5508726348233967, + 0.07991225265120794 + ], + "q90": [ + -0.08692895196223581, + 0.4241791892128114, + 0.27564319435356965, + 3.0613416507785582, + 0.8103640644268676, + -0.0019377212457458886, + 0.0799979884460188 + ], + "q99": [ + -0.06423466663816729, + 0.43296238242924867, + 0.3180399667824616, + 3.1243694762781864, + 0.8279769774488147, + 0.018234213260534936, + 0.07999979723527649 + ] + }, + "timestamp": { + "min": [ + 0.0 + ], + "max": [ + 2.5 + ], + "mean": [ + 1.0747710622710622 + ], + "std": [ + 0.6436733828410123 + ], + "count": [ + 1092 + ], + "q01": [ + -1.0000000000000002e-10 + ], + "q10": [ + 0.19774486988174178 + ], + "q50": [ + 1.0718439285711645 + ], + "q90": [ + 1.9517972546603828 + ], + "q99": [ + 2.1493516413087743 + ] + }, + "observation.images.wrist": { + "min": [ + [ + [ + 0.0 + ] + ], + [ + [ + 0.0 + ] + ], + [ + [ + 0.0 + ] + ] + ], + "max": [ + [ + [ + 1.0 + ] + ], + [ + [ + 1.0 + ] + ], + [ + [ + 1.0 + ] + ] + ], + "mean": [ + [ + [ + 0.6745428680286217 + ] + ], + [ + [ + 0.4093847920152636 + ] + ], + [ + [ + 0.2736248342673908 + ] + ] + ], + "std": [ + [ + [ + 0.006637401603505537 + ] + ], + [ + [ + 0.006226898831192667 + ] + ], + [ + [ + 0.006440289562329204 + ] + ] + ], + "count": [ + 1092 + ], + "q01": [ + [ + [ + 0.07916657623771842 + ] + ], + [ + [ + 0.0006734599316017886 + ] + ], + [ + [ + 0.0006688347089129257 + ] + ] + ], + "q10": [ + [ + [ + 0.49967314733904006 + ] + ], + [ + [ + 0.2869806094237371 + ] + ], + [ + [ + 0.18275580738727581 + ] + ] + ], + "q50": [ + [ + [ + 0.6787185730987904 + ] + ], + [ + [ + 0.4113664471464953 + ] + ], + [ + [ + 0.25633113471209906 + ] + ] + ], + "q90": [ + [ + [ + 0.8838156678876091 + ] + ], + [ + [ + 0.5499654897450393 + ] + ], + [ + [ + 0.3618649874566347 + ] + ] + ], + "q99": [ + [ + [ + 0.999891788011989 + ] + ], + [ + [ + 0.8488695155096675 + ] + ], + [ + [ + 0.852053441085276 + ] + ] + ] + } +} \ No newline at end of file diff --git a/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/conversion/expected/lerobot__val/meta/stats_gr00t.json b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/conversion/expected/lerobot__val/meta/stats_gr00t.json new file mode 100644 index 0000000000000000000000000000000000000000..4d02a57cf8410bb80060135f9a0962d9d317aa5c --- /dev/null +++ b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/conversion/expected/lerobot__val/meta/stats_gr00t.json @@ -0,0 +1,120 @@ +{ + "__format_version": 2, + "__cache_config": { + "mode": "abs" + }, + "statistics": { + "observation.state": { + "mean": [ + -0.14554812014102936, + 0.07433967292308807, + 0.1436367779970169, + 2.700871706008911, + 0.45681115984916687, + -0.48229172825813293, + 0.06972828507423401 + ], + "std": [ + 0.07678605616092682, + 0.21988309919834137, + 0.09388109296560287, + 1.109084963798523, + 0.3012843728065491, + 0.3006882667541504, + 0.0164483655244112 + ], + "min": [ + -0.37446925044059753, + -0.29720592498779297, + 0.016940969973802567, + -3.140481948852539, + -0.10707131028175354, + -1.2001688480377197, + 0.03238735347986221 + ], + "max": [ + 0.014208495616912842, + 0.5909395217895508, + 0.3558541536331177, + 3.1415467262268066, + 0.9420600533485413, + 0.09770228713750839, + 0.07999999821186066 + ], + "q01": [ + -0.3139808464050293, + -0.25229972660541533, + 0.0188405978679657, + -3.134360270500183, + -0.04521620273590088, + -1.0301372480392457, + 0.03585549935698509 + ], + "q99": [ + -0.013631248474121185, + 0.5275326526165007, + 0.33348673582077026, + 3.1367013454437256, + 0.8922865676879882, + 0.056985778063535636, + 0.07999999821186066 + ] + }, + "action": { + "mean": [ + -0.048697348684072495, + 0.24886493384838104, + -0.16479986906051636, + -0.032609131187200546, + -0.501691997051239, + 0.44703611731529236, + 0.6793776154518127 + ], + "std": [ + 0.1079770028591156, + 0.47203660011291504, + 0.14475314319133759, + 0.2849433720111847, + 0.318120539188385, + 0.35324087738990784, + 0.6349603533744812 + ], + "min": [ + -0.6572320461273193, + -1.0, + -0.8218780755996704, + -0.7686842083930969, + -0.9920897483825684, + -0.30364304780960083, + -1.0 + ], + "max": [ + 0.30179348587989807, + 0.7795099020004272, + 0.2688712179660797, + 0.8059309720993042, + 0.5333064198493958, + 1.0, + 1.0 + ], + "q01": [ + -0.4552440994977951, + -1.0, + -0.7419710528850555, + -0.475678391456604, + -0.9305388247966766, + -0.20442348539829253, + -0.8981475746631622 + ], + "q99": [ + 0.1125871029496192, + 0.7402709639072418, + 0.22590837419033025, + 0.7005072689056396, + 0.16810789704322815, + 1.0, + 1.0 + ] + } + } +} diff --git a/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/conversion/expected/lerobot__val/meta/tasks.parquet b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/conversion/expected/lerobot__val/meta/tasks.parquet new file mode 100644 index 0000000000000000000000000000000000000000..9c55ec41283f569952a3f11813c4924db9e53c7b --- /dev/null +++ b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/conversion/expected/lerobot__val/meta/tasks.parquet @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:29b49531995b967e7b4afc52add5605c15e84ad1dd445ca06dccff69df98e5f5 +size 2351 diff --git a/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/conversion/expected/lerobot__val/task_contract.json b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/conversion/expected/lerobot__val/task_contract.json new file mode 100644 index 0000000000000000000000000000000000000000..76ac5a2d1a2e11769474504874b6e909be02e9b8 --- /dev/null +++ b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/conversion/expected/lerobot__val/task_contract.json @@ -0,0 +1,131 @@ +{ + "format": "latency_bench_starvla_h1_v1", + "task": "mikasa_intercept_grab_fast", + "robot_type": "latency_mikasa_h1", + "state_dim": 7, + "action_dim": 7, + "action_horizon": 1, + "fps": 20.0, + "state_key": "observation.state", + "action_key": "action", + "video_keys": [ + "video.top", + "video.wrist" + ], + "auxiliary": {}, + "normalization": { + "state": { + "mean": [ + -0.14554812014102936, + 0.07433967292308807, + 0.1436367779970169, + 2.700871706008911, + 0.45681115984916687, + -0.48229172825813293, + 0.06972828507423401 + ], + "std": [ + 0.07678605616092682, + 0.21988309919834137, + 0.09388109296560287, + 1.109084963798523, + 0.3012843728065491, + 0.3006882667541504, + 0.0164483655244112 + ], + "min": [ + -0.37446925044059753, + -0.29720592498779297, + 0.016940969973802567, + -3.140481948852539, + -0.10707131028175354, + -1.2001688480377197, + 0.03238735347986221 + ], + "max": [ + 0.014208495616912842, + 0.5909395217895508, + 0.3558541536331177, + 3.1415467262268066, + 0.9420600533485413, + 0.09770228713750839, + 0.07999999821186066 + ], + "q01": [ + -0.3139808464050293, + -0.25229972660541533, + 0.0188405978679657, + -3.134360270500183, + -0.04521620273590088, + -1.0301372480392457, + 0.03585549935698509 + ], + "q99": [ + -0.013631248474121185, + 0.5275326526165007, + 0.33348673582077026, + 3.1367013454437256, + 0.8922865676879882, + 0.056985778063535636, + 0.07999999821186066 + ] + }, + "action": { + "mean": [ + -0.048697348684072495, + 0.24886493384838104, + -0.16479986906051636, + -0.032609131187200546, + -0.501691997051239, + 0.44703611731529236, + 0.6793776154518127 + ], + "std": [ + 0.1079770028591156, + 0.47203660011291504, + 0.14475314319133759, + 0.2849433720111847, + 0.318120539188385, + 0.35324087738990784, + 0.6349603533744812 + ], + "min": [ + -0.6572320461273193, + -1.0, + -0.8218780755996704, + -0.7686842083930969, + -0.9920897483825684, + -0.30364304780960083, + -1.0 + ], + "max": [ + 0.30179348587989807, + 0.7795099020004272, + 0.2688712179660797, + 0.8059309720993042, + 0.5333064198493958, + 1.0, + 1.0 + ], + "q01": [ + -0.4552440994977951, + -1.0, + -0.7419710528850555, + -0.475678391456604, + -0.9305388247966766, + -0.20442348539829253, + -0.8981475746631622 + ], + "q99": [ + 0.1125871029496192, + 0.7402709639072418, + 0.22590837419033025, + 0.7005072689056396, + 0.16810789704322815, + 1.0, + 1.0 + ] + } + }, + "prompt": "Intercept the rolling ball and grasp it to stop it." +} diff --git a/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/conversion/expected/mixture.json b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/conversion/expected/mixture.json new file mode 100644 index 0000000000000000000000000000000000000000..9e02367f93643e5c2266a417c2c0d3365eb1eff1 --- /dev/null +++ b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/conversion/expected/mixture.json @@ -0,0 +1,16 @@ +{ + "train": [ + [ + "/home/ubuntu/lzj/mean-profiling/intercept/data/lerobot", + 1.0, + "latency_mikasa_h1" + ] + ], + "validation": [ + [ + "/home/ubuntu/lzj/mean-profiling/intercept/data/lerobot__val", + 1.0, + "latency_mikasa_h1" + ] + ] +} diff --git a/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/metadata.json b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/metadata.json new file mode 100644 index 0000000000000000000000000000000000000000..54d199f63e88d9359de82a93e31c1864ab3f5d89 --- /dev/null +++ b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/metadata.json @@ -0,0 +1,1114 @@ +{ + "env_id": "InterceptGrabFast-VLA-v0", + "num_episodes": 275, + "episode_lengths": [ + 41, + 37, + 50, + 47, + 47, + 41, + 46, + 45, + 43, + 39, + 41, + 41, + 43, + 48, + 45, + 45, + 43, + 45, + 43, + 49, + 47, + 44, + 45, + 45, + 51, + 39, + 44, + 43, + 51, + 41, + 45, + 47, + 41, + 49, + 43, + 45, + 43, + 47, + 47, + 45, + 45, + 43, + 41, + 49, + 43, + 45, + 53, + 43, + 41, + 45, + 45, + 39, + 45, + 49, + 43, + 43, + 45, + 45, + 39, + 45, + 43, + 39, + 49, + 41, + 45, + 53, + 45, + 43, + 47, + 43, + 41, + 49, + 43, + 46, + 53, + 43, + 47, + 43, + 43, + 41, + 52, + 48, + 47, + 41, + 39, + 47, + 45, + 51, + 49, + 39, + 41, + 49, + 44, + 49, + 45, + 43, + 47, + 41, + 39, + 47, + 41, + 49, + 51, + 43, + 43, + 43, + 45, + 41, + 43, + 47, + 43, + 49, + 44, + 45, + 47, + 44, + 45, + 49, + 41, + 44, + 41, + 43, + 43, + 43, + 43, + 43, + 43, + 49, + 41, + 41, + 45, + 43, + 53, + 39, + 43, + 48, + 43, + 47, + 49, + 43, + 37, + 47, + 43, + 46, + 43, + 46, + 47, + 43, + 47, + 39, + 47, + 43, + 45, + 39, + 47, + 49, + 39, + 39, + 49, + 43, + 47, + 45, + 45, + 48, + 43, + 41, + 47, + 43, + 49, + 49, + 41, + 39, + 45, + 43, + 50, + 43, + 45, + 37, + 39, + 45, + 47, + 49, + 51, + 47, + 43, + 43, + 46, + 41, + 39, + 41, + 43, + 45, + 47, + 43, + 43, + 51, + 39, + 47, + 41, + 47, + 43, + 45, + 41, + 52, + 47, + 43, + 43, + 45, + 41, + 49, + 43, + 49, + 47, + 45, + 49, + 49, + 47, + 39, + 45, + 45, + 41, + 41, + 45, + 43, + 45, + 41, + 41, + 41, + 39, + 41, + 41, + 45, + 45, + 43, + 47, + 45, + 47, + 43, + 47, + 49, + 43, + 41, + 43, + 37, + 43, + 41, + 41, + 43, + 45, + 47, + 47, + 47, + 55, + 43, + 43, + 39, + 45, + 39, + 43, + 49, + 44, + 45, + 47, + 41, + 47, + 45, + 43, + 49, + 39, + 39, + 45, + 47, + 41, + 47, + 39 + ], + "success_once": [ + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true, + true + ], + "reward_sums": [ + 8.892977714538574, + 7.88288688659668, + 9.412217140197754, + 9.409232139587402, + 8.923410415649414, + 8.789350509643555, + 9.359082221984863, + 10.381694793701172, + 8.9065580368042, + 8.080547332763672, + 9.238605499267578, + 8.887856483459473, + 8.475488662719727, + 9.800673484802246, + 7.751443862915039, + 8.546211242675781, + 8.828935623168945, + 9.363005638122559, + 8.159473419189453, + 11.346901893615723, + 9.675544738769531, + 9.604511260986328, + 10.454998970031738, + 8.203414916992188, + 10.985596656799316, + 9.635343551635742, + 8.346609115600586, + 8.16788101196289, + 12.233668327331543, + 10.364034652709961, + 10.854011535644531, + 7.47733211517334, + 10.284761428833008, + 9.365133285522461, + 8.611555099487305, + 8.529865264892578, + 8.672831535339355, + 11.2753324508667, + 8.970187187194824, + 8.208060264587402, + 8.97749137878418, + 8.857694625854492, + 8.4456205368042, + 8.57451057434082, + 7.405007839202881, + 8.505057334899902, + 12.577935218811035, + 7.84575080871582, + 8.515436172485352, + 8.968729972839355, + 10.842937469482422, + 8.092718124389648, + 8.55148696899414, + 8.677330017089844, + 10.861316680908203, + 8.917747497558594, + 8.533430099487305, + 8.874360084533691, + 8.830503463745117, + 10.884847640991211, + 8.07717514038086, + 8.046730041503906, + 9.809685707092285, + 8.086063385009766, + 8.977666854858398, + 11.428434371948242, + 8.191499710083008, + 8.129722595214844, + 10.952189445495605, + 8.498960494995117, + 8.14847183227539, + 8.231287002563477, + 8.081901550292969, + 9.774286270141602, + 11.514646530151367, + 8.168581008911133, + 8.959518432617188, + 8.1592378616333, + 8.217885971069336, + 8.425567626953125, + 10.221562385559082, + 9.832710266113281, + 9.389862060546875, + 8.728996276855469, + 8.42551040649414, + 8.97703742980957, + 7.848455429077148, + 8.63687801361084, + 9.031641006469727, + 8.410921096801758, + 9.59266471862793, + 9.793882369995117, + 8.823936462402344, + 8.998924255371094, + 11.191147804260254, + 8.918851852416992, + 8.950267791748047, + 8.875429153442383, + 8.768182754516602, + 8.217281341552734, + 8.128216743469238, + 10.97981071472168, + 11.307100296020508, + 8.961462020874023, + 8.498649597167969, + 8.122061729431152, + 8.204160690307617, + 8.083967208862305, + 8.954179763793945, + 8.970466613769531, + 10.38818073272705, + 12.199295043945312, + 8.97357177734375, + 8.480033874511719, + 8.997074127197266, + 8.918909072875977, + 8.902657508850098, + 10.946880340576172, + 8.849684715270996, + 8.529777526855469, + 10.242470741271973, + 8.972196578979492, + 9.696046829223633, + 8.52467155456543, + 9.60664176940918, + 8.523391723632812, + 8.982571601867676, + 9.072511672973633, + 8.892273902893066, + 8.785362243652344, + 8.520249366760254, + 8.531267166137695, + 11.416980743408203, + 8.023359298706055, + 9.58596134185791, + 9.003728866577148, + 8.192112922668457, + 8.559354782104492, + 14.779024124145508, + 9.719480514526367, + 8.727720260620117, + 9.340614318847656, + 9.236281394958496, + 9.80897331237793, + 8.583600997924805, + 9.045393943786621, + 8.727560043334961, + 8.49899959564209, + 9.39169692993164, + 8.824126243591309, + 8.951570510864258, + 8.482479095458984, + 8.523787498474121, + 9.994926452636719, + 9.789587020874023, + 8.984838485717773, + 8.782999038696289, + 8.125654220581055, + 11.29443359375, + 8.052051544189453, + 8.560420989990234, + 8.57132339477539, + 9.595429420471191, + 10.56120491027832, + 8.979769706726074, + 8.456537246704102, + 8.24409294128418, + 8.519569396972656, + 9.008519172668457, + 9.093876838684082, + 7.650932312011719, + 8.431344985961914, + 8.146347045898438, + 8.181321144104004, + 9.109539985656738, + 8.496376991271973, + 8.595174789428711, + 7.9894185066223145, + 8.057146072387695, + 8.210878372192383, + 8.53508472442627, + 15.242013931274414, + 9.848501205444336, + 9.759037971496582, + 8.93541145324707, + 8.501548767089844, + 9.751532554626465, + 8.478347778320312, + 10.16085433959961, + 8.939860343933105, + 8.888265609741211, + 8.177536010742188, + 9.38625431060791, + 8.422812461853027, + 8.849225044250488, + 10.657587051391602, + 8.896005630493164, + 9.019929885864258, + 8.51335334777832, + 8.193994522094727, + 9.301592826843262, + 8.173934936523438, + 8.874101638793945, + 9.855046272277832, + 7.508752822875977, + 8.99341869354248, + 8.51372241973877, + 8.934932708740234, + 8.819690704345703, + 11.303918838500977, + 8.646509170532227, + 10.980076789855957, + 7.479002952575684, + 10.559199333190918, + 8.998167991638184, + 10.496103286743164, + 9.033164978027344, + 8.460996627807617, + 8.906722068786621, + 8.518223762512207, + 9.552021026611328, + 8.49902057647705, + 10.805758476257324, + 8.218069076538086, + 14.930100440979004, + 9.263382911682129, + 8.80577278137207, + 8.804271697998047, + 8.730260848999023, + 8.115966796875, + 7.731543064117432, + 8.555465698242188, + 8.536657333374023, + 9.933286666870117, + 8.582874298095703, + 8.462884902954102, + 10.894489288330078, + 9.19791030883789, + 9.00412368774414, + 9.062813758850098, + 8.484842300415039, + 8.089601516723633, + 8.176942825317383, + 7.602810859680176, + 8.537178039550781, + 8.801538467407227, + 8.804023742675781, + 8.197819709777832, + 8.957082748413086, + 10.911252975463867, + 10.948500633239746, + 12.005812644958496, + 11.491374015808105, + 7.537558078765869, + 7.707664489746094, + 8.812128067016602, + 8.533111572265625, + 7.975378036499023, + 10.413677215576172, + 11.417947769165039, + 10.768226623535156, + 8.855576515197754, + 9.007559776306152, + 9.653735160827637, + 8.227901458740234, + 8.068707466125488, + 9.633874893188477, + 9.974815368652344, + 9.54748821258545, + 8.11349105834961, + 9.26374626159668, + 8.592830657958984, + 7.712718486785889, + 8.670660972595215, + 7.960550785064697 + ], + "episode_seeds": [ + 0, + 1, + 2, + 3, + 4, + 5, + 6, + 7, + 8, + 9, + 10, + 11, + 12, + 13, + 14, + 15, + 16, + 17, + 18, + 19, + 100, + 101, + 102, + 103, + 104, + 105, + 106, + 107, + 108, + 109, + 110, + 111, + 112, + 113, + 114, + 115, + 116, + 117, + 118, + 119, + 120, + 121, + 122, + 123, + 124, + 125, + 126, + 127, + 128, + 129, + 130, + 131, + 132, + 133, + 134, + 135, + 136, + 137, + 138, + 139, + 140, + 141, + 142, + 143, + 144, + 145, + 146, + 147, + 148, + 149, + 150, + 151, + 152, + 153, + 154, + 155, + 156, + 157, + 158, + 159, + 160, + 161, + 162, + 163, + 164, + 165, + 166, + 167, + 168, + 169, + 170, + 171, + 172, + 173, + 174, + 175, + 176, + 177, + 178, + 179, + 180, + 181, + 182, + 183, + 184, + 185, + 186, + 187, + 188, + 189, + 190, + 191, + 192, + 193, + 194, + 195, + 196, + 197, + 198, + 199, + 20, + 21, + 22, + 23, + 24, + 25, + 26, + 27, + 28, + 29, + 200, + 201, + 202, + 203, + 204, + 205, + 206, + 207, + 208, + 209, + 210, + 211, + 212, + 213, + 214, + 215, + 216, + 217, + 218, + 219, + 220, + 221, + 222, + 223, + 224, + 225, + 226, + 227, + 228, + 229, + 230, + 231, + 232, + 233, + 234, + 235, + 236, + 237, + 238, + 239, + 240, + 241, + 242, + 243, + 244, + 245, + 246, + 247, + 248, + 249, + 250, + 251, + 252, + 253, + 254, + 255, + 256, + 257, + 258, + 259, + 260, + 261, + 262, + 263, + 264, + 265, + 266, + 267, + 268, + 269, + 270, + 271, + 272, + 273, + 274, + 30, + 31, + 32, + 33, + 34, + 35, + 36, + 37, + 38, + 39, + 40, + 41, + 42, + 43, + 44, + 45, + 46, + 47, + 48, + 49, + 50, + 51, + 52, + 53, + 54, + 55, + 56, + 57, + 58, + 59, + 60, + 61, + 62, + 63, + 64, + 65, + 66, + 67, + 68, + 69, + 70, + 71, + 72, + 73, + 74, + 75, + 76, + 77, + 78, + 79, + 80, + 81, + 82, + 83, + 84, + 85, + 86, + 87, + 88, + 89, + 90, + 91, + 92, + 93, + 94, + 95, + 96, + 97, + 98, + 99 + ], + "batched_source_dir": "/home/ubuntu/lzj/mean-profiling/intercept/data/data_npz/_batched/intercept_grab_fast_vla_v0", + "unbatched_dir": "/home/ubuntu/lzj/mean-profiling/intercept/data/data_npz/intercept_grab_fast_vla_v0" +} \ No newline at end of file diff --git a/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_0.npz b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_0.npz new file mode 100644 index 0000000000000000000000000000000000000000..abe79640f8a816b46997addaa39de1c9b3d048d4 --- /dev/null +++ b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_0.npz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:7f46ff64233ec31a56a61dae27c793633d13c13781b031ce8da94a6a113b5448 +size 4044802 diff --git a/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_1.npz b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_1.npz new file mode 100644 index 0000000000000000000000000000000000000000..2970d3e33d7c3d02b4a073e79c11976707d6c0aa --- /dev/null +++ b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_1.npz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6d743ab66ad0541b61d462bef9e6a062ab02a52974a7f8f62c0d38596a51944b +size 3651070 diff --git a/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_10.npz b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_10.npz new file mode 100644 index 0000000000000000000000000000000000000000..daacee840d57157bb105013b1a1f37810b95837a --- /dev/null +++ b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_10.npz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:348a02b483ec901d6db5d7e50ab5838bdfb08776b2e8f5e2abe80b7b4936a856 +size 4044802 diff --git a/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_100.npz b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_100.npz new file mode 100644 index 0000000000000000000000000000000000000000..9e3e1ce03bd363814cacf68f36c4d7948bb42039 --- /dev/null +++ b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_100.npz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:15e93b5d927bacfb43d309989d7475af98e1a2f03753542fc40c002b54f7c3be +size 4044802 diff --git a/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_101.npz b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_101.npz new file mode 100644 index 0000000000000000000000000000000000000000..5de13f53fe942c24a0e1be0c3d45340ff76aa091 --- /dev/null +++ b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_101.npz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:525af0c145e326d3064c35e6fd9fc5b9bedc2ebf8fceaa0031b5787094cee260 +size 4832266 diff --git a/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_102.npz b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_102.npz new file mode 100644 index 0000000000000000000000000000000000000000..3df9211be2a9cb56d2388ec84e36432ac46e90a6 --- /dev/null +++ b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_102.npz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9e30c999efee7e7668317400ce1926f2d80ee7d17a44ea30b04cfa0a4c3ebff6 +size 5029132 diff --git a/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_103.npz b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_103.npz new file mode 100644 index 0000000000000000000000000000000000000000..4847d04a8d05880290da2c42feb633d1a332fa37 --- /dev/null +++ b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_103.npz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a483ae377f694ddecbe35420f2c9759188e87cbc493247507c702b8be47bd38c +size 4241668 diff --git a/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_104.npz b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_104.npz new file mode 100644 index 0000000000000000000000000000000000000000..203a3d76265d0d6d9085782a29104ba2f8cd25ca --- /dev/null +++ b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_104.npz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:41b706b4e791b9ef7d2e2535908cb9a072978d955be48a23ef408bdf53d0039e +size 4241668 diff --git a/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_105.npz b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_105.npz new file mode 100644 index 0000000000000000000000000000000000000000..6b93c5b3f543941909e3fab9bb71f00b116ff56d --- /dev/null +++ b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_105.npz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1c361ec3fddf99a8b45736b72248b972460c5b1d479de1ea0341573f3d388901 +size 4241668 diff --git a/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_106.npz b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_106.npz new file mode 100644 index 0000000000000000000000000000000000000000..a3605060b88dee4644e5425d0d563f40f95777ac --- /dev/null +++ b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_106.npz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:61b9059cd8ead3031aa02b778c8e343362a1118cebe1202bf71f77e57c7ddadd +size 4438534 diff --git a/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_107.npz b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_107.npz new file mode 100644 index 0000000000000000000000000000000000000000..dbaae8d8e0c57dad3bab2868eec0f6dd77f28ed3 --- /dev/null +++ b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_107.npz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2a68f80ccccba23fb92c97061e9100594a9d13c0408b65e1d91b455731cb9176 +size 4044802 diff --git a/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_108.npz b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_108.npz new file mode 100644 index 0000000000000000000000000000000000000000..0bbb07853ced9dbe4a28f07c5b6f5ab7b7c39ce3 --- /dev/null +++ b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_108.npz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ed52a51d6092b34ed9eeef52af0a5299861ead967d2ab15c1d0d8634d908934f +size 4241668 diff --git a/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_109.npz b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_109.npz new file mode 100644 index 0000000000000000000000000000000000000000..ed8ee4f6acf56e9455d519e5bc9e09ee1a28dcc2 --- /dev/null +++ b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_109.npz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:49c02fd4d1bdd82f88b641deccf31c350eb7acc05077af55523983c53689c1c0 +size 4635400 diff --git a/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_11.npz b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_11.npz new file mode 100644 index 0000000000000000000000000000000000000000..ce0b204687665902ff658c766aa6fda766008046 --- /dev/null +++ b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_11.npz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3627f87b6d9c7544d75c39310ca7a53042cee43dd18e1e24388ff1b1531c9843 +size 4044802 diff --git a/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_110.npz b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_110.npz new file mode 100644 index 0000000000000000000000000000000000000000..626821f66897c173cd9f3734a606e2dc798196a5 --- /dev/null +++ b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_110.npz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:5c555b4ade1b2f3a22faaa0dde6bdabf83a194208a6228d27ad46604746aa3be +size 4241668 diff --git a/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_111.npz b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_111.npz new file mode 100644 index 0000000000000000000000000000000000000000..323b3117186f0080a1768fb739b082c1d8eb1cb9 --- /dev/null +++ b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_111.npz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:10a164d5b3b766aff39e86300951e0a3409ffb80eddeb84389ad4e60958e6af7 +size 4832266 diff --git a/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_112.npz b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_112.npz new file mode 100644 index 0000000000000000000000000000000000000000..14cc78481b1c2b040cb802b313d04c036432fbbd --- /dev/null +++ b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_112.npz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:bb632da6917093027bcfc388570f5520b0166806b74edc147aec0754da29737f +size 4340101 diff --git a/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_113.npz b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_113.npz new file mode 100644 index 0000000000000000000000000000000000000000..1ca0e61b9f2a7f41a6c448dda1019972422d2592 --- /dev/null +++ b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_113.npz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:18d449d81ac535f9271228b5b0fe4316b52a6fa1d8ff1381ed81e5f9cf29a86c +size 4438534 diff --git a/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_114.npz b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_114.npz new file mode 100644 index 0000000000000000000000000000000000000000..8d1eff6689f90baa7c7acccbfa17c2856842a01c --- /dev/null +++ b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_114.npz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d3fb11a5b98307f3abcf504ff6d5d7798f5609ed303c4301da50fe30ebb0b99d +size 4635400 diff --git a/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_115.npz b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_115.npz new file mode 100644 index 0000000000000000000000000000000000000000..1d6778c7e1b862df6195e362e8e0e40fdd81c15e --- /dev/null +++ b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_115.npz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2f1855433bbfc6254f67635201acacdf3df8d465632804c05b8c6177cffba706 +size 4340101 diff --git a/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_116.npz b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_116.npz new file mode 100644 index 0000000000000000000000000000000000000000..b2b5424d23fe7cb9fc52899bfb90a67678e46563 --- /dev/null +++ b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_116.npz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:72fb4d40b458d86b6ffb4bf2fb12bc92d16e857adccc7e56e17dfc05308db356 +size 4438534 diff --git a/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_117.npz b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_117.npz new file mode 100644 index 0000000000000000000000000000000000000000..78afe80e17f0245e4f771a254e3c0bcb435815e1 --- /dev/null +++ b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_117.npz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c0c7ad42cc4c5826b9150507f0c0e045e1eebf278452d7dd3635a73a18e19b43 +size 4832266 diff --git a/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_118.npz b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_118.npz new file mode 100644 index 0000000000000000000000000000000000000000..a17caf2edb7337b38447b21ae88de7fca6fdde3b --- /dev/null +++ b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_118.npz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1404382f5404b2118223f013835f8f5cc8600d3815961223fbfebd2ae558f935 +size 4044802 diff --git a/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_119.npz b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_119.npz new file mode 100644 index 0000000000000000000000000000000000000000..fadc07cd922169c7dcdb7dadc009f72503c35d9a --- /dev/null +++ b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_119.npz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:115d5be6bbb4311a2f30fbf081d651a39dbb1db2e8b958698db4bc183437f1c3 +size 4340101 diff --git a/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_12.npz b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_12.npz new file mode 100644 index 0000000000000000000000000000000000000000..f385fbcee04ce6a61a2762f99f76f3ab239acbaa --- /dev/null +++ b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_12.npz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2056d656e52d4763c76e376e80239889e6100d7469d15001da7129b7f78fa684 +size 4241668 diff --git a/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_120.npz b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_120.npz new file mode 100644 index 0000000000000000000000000000000000000000..7fcd859b09d2289835792aba9acd033edac2155d --- /dev/null +++ b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_120.npz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f53b257fa2de0a710087543a54941b0fefb18106333e7139e70b1cfddd59be96 +size 4044802 diff --git a/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_121.npz b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_121.npz new file mode 100644 index 0000000000000000000000000000000000000000..75817957811d33a88fa34dea3d1d1faef7d35528 --- /dev/null +++ b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_121.npz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9406e14ae0c52b80ac95e9acd2a40a4cf7cc5e8edb7a7b8b1c41e33469125322 +size 4241668 diff --git a/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_122.npz b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_122.npz new file mode 100644 index 0000000000000000000000000000000000000000..f85bc58c9b6e22f1d1957eb23f4dd9c21f416131 --- /dev/null +++ b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_122.npz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e9696baa716f38fa519895dc9df10c4780c606382b2a7c00494cd110ddc4df2c +size 4241668 diff --git a/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_123.npz b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_123.npz new file mode 100644 index 0000000000000000000000000000000000000000..009f6aeb7f41494d8d8651a8cde46af22ea466ee --- /dev/null +++ b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_123.npz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:7bf61acc286c1a830ebbcedef9104b7835d862471e00a4fc55cef5bebe8a26c0 +size 4241668 diff --git a/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_124.npz b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_124.npz new file mode 100644 index 0000000000000000000000000000000000000000..0f131864e4fdd24e3f6f5d5786a6e3cd04644faa --- /dev/null +++ b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_124.npz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6897861a7a58d5b3f5d3b8c9e9cb73345eb61723d920f74ecb66b773e5cef6fc +size 4241668 diff --git a/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_125.npz b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_125.npz new file mode 100644 index 0000000000000000000000000000000000000000..119d95e28a03af45476af211d9f9b1abdbe8abe0 --- /dev/null +++ b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_125.npz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:059d4ec1d5380c1fb70440d3ceddca68285eb777c26c133cdf3a1b347b398c60 +size 4241668 diff --git a/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_126.npz b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_126.npz new file mode 100644 index 0000000000000000000000000000000000000000..7c02b6fdd60a7d46aad50e309996c72fd5710e33 --- /dev/null +++ b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_126.npz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3a3f8db5089c6b64fda6621e234c2a15afb3787274c0b82c61860a91cf9a6e9d +size 4241668 diff --git a/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_127.npz b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_127.npz new file mode 100644 index 0000000000000000000000000000000000000000..dff941538c970af6fc7d6ef0601ec14427c30d98 --- /dev/null +++ b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_127.npz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:246fcbc029d51882633c18ce0fd6406e3a24a35ed9737984b4ff95fcec12d6dd +size 4832266 diff --git a/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_128.npz b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_128.npz new file mode 100644 index 0000000000000000000000000000000000000000..c480ffecc1a60a8213be17b7b720503b4f011f69 --- /dev/null +++ b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_128.npz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c7289651b9f991aa868b3316071b954f8fc4769893fb66a089f5ab74991b010f +size 4044802 diff --git a/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_129.npz b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_129.npz new file mode 100644 index 0000000000000000000000000000000000000000..be38e76ead5ee78a2c0c664e184ef0b06547660f --- /dev/null +++ b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_129.npz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1bbb18a9fda56d8997ce76ed03bc43532007d2a1c337d6ac662332c419ab4ee1 +size 4044802 diff --git a/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_13.npz b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_13.npz new file mode 100644 index 0000000000000000000000000000000000000000..14dc50e733b461ce2280668b1b1833beaae04c30 --- /dev/null +++ b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_13.npz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:110f8907a57320f59e9ba4b5b821ad8f841d2b2e59cde1b646bc7c3c2e57ed19 +size 4733833 diff --git a/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_130.npz b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_130.npz new file mode 100644 index 0000000000000000000000000000000000000000..26f7d68c7f3302f3e44c8934864b4fab035b31d5 --- /dev/null +++ b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_130.npz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d78ea6f511245fddc4bd402956559ea20215a6d2dfc1beca8d493e83de7c71bc +size 4438534 diff --git a/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_131.npz b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_131.npz new file mode 100644 index 0000000000000000000000000000000000000000..d6137811903468793ed8e152630f54f54eb1ab47 --- /dev/null +++ b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_131.npz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4df76f00f851c05302cdd6018f2d9d661c89a94e4f4078eeb376a43a5b8efd99 +size 4241668 diff --git a/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_132.npz b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_132.npz new file mode 100644 index 0000000000000000000000000000000000000000..bd42e645fb47e2d3b5a7a43cdd5199f65bd81946 --- /dev/null +++ b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_132.npz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:8c229ccaaffb0d18319a237873188d34e98bd6985d0013ff35dbf1c6201b79a1 +size 5225998 diff --git a/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_133.npz b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_133.npz new file mode 100644 index 0000000000000000000000000000000000000000..5326b8ec2568455bf89fa7814bf938f72040849f --- /dev/null +++ b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_133.npz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:73a3e1cbae75db6ba339f0d403e663cbf87a03a75150ffaa287c70af3a8abb16 +size 3847936 diff --git a/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_134.npz b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_134.npz new file mode 100644 index 0000000000000000000000000000000000000000..68bc5b7443f03143a8ce467008dbfc45dcd204df --- /dev/null +++ b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_134.npz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:0c1f3170be76488b1f77bd2cd023da14072098939a5457fecce18c578764ae17 +size 4241668 diff --git a/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_135.npz b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_135.npz new file mode 100644 index 0000000000000000000000000000000000000000..a9f4aba69d12a507d6e4d9562debbeefad505913 --- /dev/null +++ b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_135.npz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1ef9d0a7c89cb7f402c33b3660765a4ddf701efeafd5f9f689afc2e8182255e6 +size 4733833 diff --git a/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_136.npz b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_136.npz new file mode 100644 index 0000000000000000000000000000000000000000..a8ff6dbbea022a135b8170154dae00892ee50c8f --- /dev/null +++ b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_136.npz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:86b6575de22724f342d3661809199a39297a191965bd62c811dcb97ccc377986 +size 4241668 diff --git a/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_137.npz b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_137.npz new file mode 100644 index 0000000000000000000000000000000000000000..23ae9b0faa47d4bc9a729f89a9d90ff10314c2ce --- /dev/null +++ b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_137.npz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a351d9eeef529fc28ac6f01d0895a052914d8bc44d3be593c35d88702eaf6056 +size 4635400 diff --git a/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_138.npz b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_138.npz new file mode 100644 index 0000000000000000000000000000000000000000..ecbed0b8a20174e5e73c318f89ef65ee7195bdbe --- /dev/null +++ b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_138.npz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3bb17983e272681485f473735132fd555cd595ea7a9385c5dd45c8159d327afc +size 4832266 diff --git a/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_139.npz b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_139.npz new file mode 100644 index 0000000000000000000000000000000000000000..3fc2d79de1f533248814d262315700396f1fd9de --- /dev/null +++ b/latency-aware/intercept-grab-fast/mean_rtx3090_h1_s123_275ep_v1/demonstrations/raw/raw_npz/train_data_139.npz @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:5b2cbefbb1d1e164ebb0fb8d750b4a8f00b05e6e74de037325881e3b292df2c3 +size 4241668