jellyho/aloha_lift_box

Robotics Dataset

A LeRobot-format dataset of 50 episodes (11,572 frames) of a simulated robot lifting a box, captured from three RGB cameras (agentview, left and right wrist) with joint and end-effector state/action data.

Catalog metadata

Repository owner: jellyho
Source: Hugging Face / LeRobot
Episodes: 50
Frames: 11572
Cameras: 3
Frames per second: 20
License: apache-2.0

Browse the catalog to compare robots, tasks, sensors, and licenses. Sign in to open protected download and model links.