Datasets:
metadata
license: other
tags:
- robotics
- lerobot
- manipulation
- exumi
- nmotion-initiative
Pick Brocolli
Visualizers
Overview
This dataset is a conversion from an upstream robotics dataset into LeRobot v3-compatible format. It is intended to provide reproducible access in a unified schema.
Source Dataset
- Dataset ID:
exumi-pick-brocolli - Project:
exUMI - Task:
Pick Brocolli - Upstream download/source URL: https://huggingface.co/datasets/silicx/exUMI/resolve/main/pick_brocolli.zarr.zip?download=true
- Upstream project page: https://silicx.github.io/exUMI/
- Contact: silicxuyue@outlook.com
- Original paper: Xu et al. 2025
- Notes: 170 demos; image+tactile
Dataset Format
- Storage format: LeRobot v3 parquet-based dataset
- Total episodes: 30
- Total frames: 13137
- FPS: 30
- Robot type: generic_robot
- Observation feature:
observation.statewith shape[7] - Action feature:
actionwith shape[7] - Video streams: present
Video Modalities
observation.images.camera0_rgb: shape[224, 224, 3]- Frame rate:
30FPS
Conversion Notes
- Target schema: LeRobot v3
- Conversion path: ingest -> preprocess -> zarr probe -> convert 1 episode -> upload
actionfallback may be used when upstreamdata/actionis missing.
Dataset Structure
meta/info.json:
{
"codebase_version": "v3.0",
"robot_type": "generic_robot",
"total_episodes": 30,
"total_frames": 13137,
"fps": 30,
"splits": {"train": "0:30"},
"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.state": {"dtype": "float32", "shape": [7]},
"action": {"dtype": "float32", "shape": [7]}
}
}
Data Splits
{'train': '0:30'}
Citation
This dataset was created using LeRobot. Please also cite the original upstream project and paper listed above.
Attribution
- Original dataset credit remains with upstream creators.
- Conversion and curation contribution:
- Convertor: David Yuan, davidy02@stanford.edu, Stanford University
- NMotion Initiative (opensource community): we make embodied data accessible