rtrtwrw2026apache-2.0
granular-material-pour — buckwheat, forward
439 teleoperated buckwheat pouring episodes on a YAM arm where a policy must read the scale to determine target weight, recorded at 30 Hz with four cameras and a load cell.
Downloads92
Episodes439
Hours6.0
Why This Matters for Physical AI
This dataset demonstrates state-conditioned instruction following for continuous manipulation with sensorimotor feedback, where policies must learn to read real-time sensor signals rather than rely on discrete language specifications.
Technical Profile
- Modalities
- rgbforce_torqueproprioception
- Robot Embodiments
- YAM arm
- Action Space
- joint_positions
- Environment
- lab
- Task Types
- pouringmanipulation
- Episodes
- 439
- Total Hours
- 6.0
- Data Format
- LeRobot
- Annotation Types
- language_instructionsstate_labels
- License
- apache-2.0
Access
Need custom rgb data?
Claru builds purpose-built datasets for lab applications with dense human annotations and quality assurance.
Request a Sample Pack