XvKuoMingapache-2.0
so101_chess_corrections
One thousand expert corrections for an SO-101 arm capturing chess pieces, consisting of episodes where a trained policy failed and an expert demonstrated the correct recovery. Recorded in MuJoCo simulation.
Downloads37
Episodes1,000
Why This Matters for Physical AI
This dataset demonstrates the application of human expert corrections to improve robotic manipulation policy performance through DAGGER-style learning, showing how recovery demonstrations can address specific failure modes in learned behaviors.
Technical Profile
- Modalities
- rgbproprioception
- Robot Embodiments
- SO-101
- Action Space
- joint_positions
- Environment
- simulation
- Task Types
- manipulationgraspingpick_and_place
- Episodes
- 1,000
- Data Format
- LeRobot
- Annotation Types
- language_instructionsaction_labels
- License
- apache-2.0
Access
Need custom rgb data?
Claru builds purpose-built datasets for simulation applications with dense human annotations and quality assurance.
Request a Sample Pack