Buckets:

ayushexel's picture
|
download
raw
933 Bytes
metadata
license: apache-2.0
task_categories:
  - robotics
tags:
  - LeRobot
  - lance
  - lancedb

pusht, stored in Lance

lerobot/pusht converted to the Lance layout that LeRobot's LanceDBDataset reads natively. Same meta/ sidecar as the original, tabular features in frames.lance, encoded videos in a videos.lance blob v2 column.

from lerobot.datasets import LeRobotDataset  # not needed, shown for contrast
from lerobot.datasets import LanceDBDataset

ds = LanceDBDataset("lance-format/pusht-lance")  # streams tables from the Hub, only meta/ is downloaded
item = ds[0]

Or just point training at it. The dataset factory detects the format automatically:

lerobot-train --dataset.repo_id=lance-format/pusht-lance ...

Requires lancedb>=0.35.0b2 (pip install --pre --extra-index-url https://pypi.fury.io/lancedb/ lancedb) and the lance-dataloader branch of LeRobot until the upstream PR lands.

Xet Storage Details

Size:
933 Bytes
·
Xet hash:
0b0afc5df545479a69597a02f4ddda4ac0aba94812dd7a6d63415db5e3eead5f

Xet efficiently stores files, intelligently splitting them into unique chunks and accelerating uploads and downloads. More info.