Datasets:
This dataset was created using LeRobot.
Dataset Description
LeRobot copy of source-n300-seed0 from Autobrik/ecse552-ocbench-block
(revision of the exact .npz data and its documentation), for the McGill ECSE 552 project
Learning to Manipulate a New Object from Action-Free Videos
(code).
Scripted-expert episodes in OCBench (UR5e + Robotiq, MuJoCo): object source (red 6x6x6 cm cube),
300 episodes, 10 Hz, three 224 x 224 cameras (front, side, wrist), 7 normalized actions
(six joint increments, one gripper increment). The cube starts in a 20 x 20 cm area beside a
fixed goal marked by a 12 cm gray square; the arm always starts from the same pose. Success:
the cube rests inside the square, released and still, for 1 s.
observation.state (joints) and observation.environment_state (cube pose) are privileged
simulator state: the project's models use images only. Episode i here is
episode_{i:05d}.npz in the source collection (seed in its manifest.json).
Videos are lossy (AV1); train from the .npz copy for exact pixels.
- Homepage: [More Information Needed]
- Paper: [More Information Needed]
- License: mit
Dataset Structure
{
"codebase_version": "v3.0",
"fps": 10,
"features": {
"observation.images.front": {
"dtype": "video",
"shape": [
224,
224,
3
],
"names": [
"height",
"width",
"channels"
],
"info": {
"video.height": 224,
"video.width": 224,
"video.codec": "av1",
"video.pix_fmt": "yuv420p",
"video.fps": 10,
"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.side": {
"dtype": "video",
"shape": [
224,
224,
3
],
"names": [
"height",
"width",
"channels"
],
"info": {
"video.height": 224,
"video.width": 224,
"video.codec": "av1",
"video.pix_fmt": "yuv420p",
"video.fps": 10,
"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.wrist": {
"dtype": "video",
"shape": [
224,
224,
3
],
"names": [
"height",
"width",
"channels"
],
"info": {
"video.height": 224,
"video.width": 224,
"video.codec": "av1",
"video.pix_fmt": "yuv420p",
"video.fps": 10,
"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.state": {
"dtype": "float32",
"shape": [
7
],
"names": [
"shoulder_pan",
"shoulder_lift",
"elbow",
"wrist_1",
"wrist_2",
"wrist_3",
"gripper_driver"
]
},
"observation.environment_state": {
"dtype": "float32",
"shape": [
7
],
"names": [
"cube_x",
"cube_y",
"cube_z",
"cube_qw",
"cube_qx",
"cube_qy",
"cube_qz"
]
},
"action": {
"dtype": "float32",
"shape": [
7
],
"names": [
"d_shoulder_pan",
"d_shoulder_lift",
"d_elbow",
"d_wrist_1",
"d_wrist_2",
"d_wrist_3",
"d_gripper"
]
},
"next.reward": {
"dtype": "float32",
"shape": [
1
],
"names": null
},
"next.done": {
"dtype": "bool",
"shape": [
1
],
"names": null
},
"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
}
},
"total_episodes": 300,
"total_frames": 24568,
"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": "ur5e_robotiq",
"splits": {
"train": "0:300"
}
}
Citation
BibTeX:
[More Information Needed]
- Downloads last month
- 47