4k episodes · 30 fps · 3 cameras · 640×360 av1
This dataset was created using LeRobot.
Note on normalization statistics
On 2026-10-17 the quantile entries (q01, q10, q50, q90, q99) in meta/stats.json were replaced
for all numeric, non-index features (observation.state, action and the labels.* features).
LeRobot's recompute_stats aggregates quantiles as the frame-count-weighted mean of per-episode quantiles. With
90-frame episodes this gives quantiles far narrower than the data: under QUANTILES normalization (the pi0.5 default)
37-92% of frames per dimension fell outside [-1, 1]. The new values are exact quantiles over all 360,000 frames, so
about 2% of frames per dimension fall outside [-1, 1].
min, max, mean, std, count and the image stats are unchanged. The action stats are for absolute actions;
recompute them (and fix the quantiles again) before training with relative actions. The episode_index and index
stats were not touched and are not used for training. Previous stats.json: commit 112d8d5.
Dataset Structure
{
"codebase_version": "v3.0",
"fps": 30,
"features": {
"observation.images.middle_left_camera": {
"dtype": "video",
"shape": [
3,
360,
640
],
"names": [
"channels",
"height",
"width"
],
"info": {
"video.height": 360,
"video.width": 640,
"video.codec": "av1",
"video.pix_fmt": "yuv420p",
"video.fps": 30,
"video.channels": 3,
"has_audio": false,
"video.g": 2,
"video.crf": 30,
"video.preset": 12,
"video.fast_decode": 0,
"video.video_backend": "pyav",
"video.extra_options": {},
"is_depth_map": false
}
},
"observation.images.top_center_camera": {
"dtype": "video",
"shape": [
3,
360,
640
],
"names": [
"channels",
"height",
"width"
],
"info": {
"video.height": 360,
"video.width": 640,
"video.codec": "av1",
"video.pix_fmt": "yuv420p",
"video.fps": 30,
"video.channels": 3,
"has_audio": false,
"video.g": 2,
"video.crf": 30,
"video.preset": 12,
"video.fast_decode": 0,
"video.video_backend": "pyav",
"video.extra_options": {},
"is_depth_map": false
}
},
"observation.images.top_left_camera": {
"dtype": "video",
"shape": [
3,
360,
640
],
"names": [
"channels",
"height",
"width"
],
"info": {
"video.height": 360,
"video.width": 640,
"video.codec": "av1",
"video.pix_fmt": "yuv420p",
"video.fps": 30,
"video.channels": 3,
"has_audio": false,
"video.g": 2,
"video.crf": 30,
"video.preset": 12,
"video.fast_decode": 0,
"video.video_backend": "pyav",
"video.extra_options": {},
"is_depth_map": false
}
},
"labels.delta_robot_pose": {
"dtype": "float32",
"shape": [
6
],
"names": [
"delta_robot_pose.x",
"delta_robot_pose.y",
"delta_robot_pose.z",
"delta_robot_pose.rx",
"delta_robot_pose.ry",
"delta_robot_pose.rz"
]
},
"labels.hand_pose": {
"dtype": "float32",
"shape": [
9
],
"names": [
"hand_pose.x",
"hand_pose.y",
"hand_pose.z",
"hand_pose.r11",
"hand_pose.r21",
"hand_pose.r31",
"hand_pose.r12",
"hand_pose.r22",
"hand_pose.r32"
]
},
"labels.target_joint_positions": {
"dtype": "float32",
"shape": [
6
],
"names": [
"target_joint_positions.0",
"target_joint_positions.1",
"target_joint_positions.2",
"target_joint_positions.3",
"target_joint_positions.4",
"target_joint_positions.5"
]
},
"action": {
"dtype": "float32",
"shape": [
9
],
"names": [
"target_robot_pose.x",
"target_robot_pose.y",
"target_robot_pose.z",
"target_robot_pose.r11",
"target_robot_pose.r21",
"target_robot_pose.r31",
"target_robot_pose.r12",
"target_robot_pose.r22",
"target_robot_pose.r32"
]
},
"labels.target_robot_pose": {
"dtype": "float32",
"shape": [
9
],
"names": [
"target_robot_pose.x",
"target_robot_pose.y",
"target_robot_pose.z",
"target_robot_pose.r11",
"target_robot_pose.r21",
"target_robot_pose.r31",
"target_robot_pose.r12",
"target_robot_pose.r22",
"target_robot_pose.r32"
]
},
"labels.pickup_error": {
"dtype": "float32",
"shape": [
7
],
"names": [
"pickup_error.x",
"pickup_error.y",
"pickup_error.z",
"pickup_error.quat_w",
"pickup_error.quat_x",
"pickup_error.quat_y",
"pickup_error.quat_z"
]
},
"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
},
"observation.state": {
"dtype": "float32",
"shape": [
9
],
"names": [
"current_robot_pose.x",
"current_robot_pose.y",
"current_robot_pose.z",
"current_robot_pose.r11",
"current_robot_pose.r21",
"current_robot_pose.r31",
"current_robot_pose.r12",
"current_robot_pose.r22",
"current_robot_pose.r32"
]
}
},
"total_episodes": 4000,
"total_frames": 360000,
"total_tasks": 1,
"chunks_size": 1000,
"data_files_size_in_mb": 100,
"video_files_size_in_mb": 200,
"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",
"robot_type": null,
"splits": {
"train": "0:4000"
}
}
Citation
BibTeX:
[More Information Needed]
- Downloads last month
- 71