Loading post…
Hugging FaceFrancesco Capuano, Michel Aractingi, Quentin Lhoest, Caroline Pascal, Pepijn Kooijmans, Jade Choghari, Remi Cadene, Simon Alibert, Adil Zouitine, Martino Russi, Steven PalmaAnnouncement
`LeRobotDataset:v3.0`: Bringing large-scale datasets to `lerobot`
Summary
LeRobotDataset:v3.0 is a new release for lerobot, Hugging Face’s robotics library, addressing file-system limitations encountered when LeRobotDataset:v2 stored one episode per file at million-episode scale. Version 3 packs multiple episodes into shared Parquet and MP4 files, while relational JSON and chunked episode metadata identify episode boundaries and locate the corresponding tabular and visual data. The format separates tabular data, visual data, and metadata; it stores low-dimensional readings in Apache Parquet, concatenates camera frames into MP4 videos, and supports native temporal windowing through delta_timestamps. It also introduces StreamingLeRobotDataset for processing batches from the Hugging Face Hub without downloading the full dataset locally, and provides a one-line conversion utility for datasets in the LeRobotDataset format. The release is planned for lerobot-v0.4.0 and is presented as a step toward datasets with millions of episodes.
Context
LeRobotDataset:v2 stored one episode per file, which created file-system limitations when datasets grew to millions of episodes. The new release is motivated by the need to keep storage manageable while supporting large robotics datasets and processing them without downloading entire collections locally.
Approach / What changed
LeRobotDataset:v3 concatenates data from multiple episodes into larger, chunked Parquet and MP4 files, using relational metadata to retrieve episode-specific boundaries and storage pointers. It separates tabular data, visual data, and metadata, adds the StreamingLeRobotDataset interface, supports temporal windowing with delta_timestamps, and provides a one-line utility for converting datasets in the LeRobotDataset format.
Takeaways
meta/info.json defines feature schemas, data types, fps, codebase version, and file path templates; meta/stats.json stores aggregate mean, standard deviation, minimum, and maximum values.
meta/tasks.jsonl maps natural-language task descriptions to integer task indices for task-conditioned policy training.
PyTorch DataLoader can collate LeRobotDataset sample dictionaries into batched tensors, including temporal image stacks when delta_timestamps is configured.
Source and freshness
- Source
- huggingface.co
- Published
- Added to Yomu
- Last updated
Summaries are generated from the original by a language model. Check the original before relying on it.
Keep reading
Related reading
The Good Stuff
Paid picks we use ourselves. Links carry a referral code — a subscription through one earns us a commission at no extra cost to you.
The best engineering reads, once a week
One useful email. Unsubscribe anytime.