windjjmixed-rights-see-rights-record
Datumwell Handwriting — Inspired by AA-Robotics v0.1
A research release of 200 synthetic writing episodes inspired by the 'Write hello world' task in AA-Robotics v0.1, packaging synchronized visual observations and joint targets in LeRobot v3 for tabletop handwriting tasks.
Downloads413
Episodes200
Likes1
Why This Matters for Physical AI
This dataset provides synchronized visual and proprioceptive data for vision-language-action model training on fine-grained manipulation tasks, demonstrating how synthetic data can support VLM/VLA research for tabletop writing and precision arm control.
Technical Profile
- Modalities
- rgbproprioceptionforce_torque
- Robot Embodiments
- NexArm
- Action Space
- joint_positions
- Environment
- simulation
- Task Types
- manipulationhandwriting
- Episodes
- 200
- Data Format
- LeRobot
- Annotation Types
- language_instructionsaction_labels
- License
- mixed-rights-see-rights-record
Access
Need custom rgb data?
Claru builds purpose-built datasets for simulation applications with dense human annotations and quality assurance.
Request a Sample Pack