DaoyuanZhu/wb_pointed_location_pick_object
Robotics Dataset
Whole-body teleoperation dataset from a Unitree G1 humanoid robot with RGB cameras, capturing tasks where a human points to a location and the robot picks up the target object. Includes 83 episodes with 65,664 frames, 6 camera views, and rich annotations such as hand and body pose landmarks.
Catalog metadata
- Repository owner
- DaoyuanZhu
- Source
- Hugging Face / LeRobot
- Robot
- Unitree_G1_WholeBody_RGB
- Episodes
- 83
- Frames
- 65664
- Cameras
- 6
- Frames per second
- 30
- License
- apache-2.0
Example tasks
- go to the place pointed by the human and pick up the target object
Browse the catalog to compare robots, tasks, sensors, and licenses. Sign in to open protected download and model links.