wantobcm/so101_box2bowl

Robotics Dataset

This dataset contains 162 episodes of a SO-101 follower robot performing a box-to-bowl placement task, with 75,251 frames of RGB video from wrist and front cameras, plus proprioceptive state and action data.

Catalog metadata

Repository owner: wantobcm
Source: Hugging Face / LeRobot
Robot: so101_follower
Episodes: 162
Frames: 75251
Cameras: 2
Frames per second: 30
License: apache-2.0

Example tasks

Grab the box and place into bowl.

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