Skip to content

Dataset · WorldEngineAI

Bimanual manipulation dataset for robotics

WorldEngineAI/WEB-Dataset

A large-scale dataset of real bimanual robot video and telemetry with text descriptions for 90 everyday tasks.

What's inside

Contains between 100k and 1M records of actions by a real bimanual YAM manipulator operated via GELLO. Data includes text annotations for 90 everyday tasks, joint states, control commands, and synchronized 60Hz video from three cameras.

What you can do with it

  • Post-training Vision-Language-Action (VLA) foundation models
  • Training bimanual manipulation policies from video streams and text instructions
Size
100K–1M rows
Downloads this month
193.1K
Likes
2
Access
on request on Hugging Face

How to load it

from datasets import load_dataset

ds = load_dataset("WorldEngineAI/WEB-Dataset")