ShareRobot: 51,403 Robotic Episodes with Affordance and Trajectory Labels

BAAI developed ShareRobot, a collection of 51,403 robotic episodes with 30 frames each, to enhance multi-dimensional robotic capabilities. The data includes labels for task planning, object affordance, and end-effector trajectories using 50 distinct prompt templates.

Use Cases

• Robotic task planning using the 10 question types derived from RoboVQA
• Object affordance detection within the 30-frame image sequences
• End-effector trajectory prediction for robotic manipulation tasks

Strengths

• 51,403 labeled episodes
• 30 frames per episode providing temporal context
• Multi-dimensional labeling covering planning, affordance, and trajectories
• 50 distinct prompt templates (5 per question type)

Limitations

• Lack of explicit column documentation in the repository metadata
• Potential geographic bias toward US-based environments as indicated by metadata tags

Provenance

Source

BAAI (Beijing Academy of Artificial Intelligence)

Collection Method

Data generation using 5 templates for each of the 10 RoboVQA question types.

Freshness

Last updated August 2025.

Geography

United States