VShirokun2026cc-by-4.0
office-rover-grasp-hd-v21
LeRobot v2.1 dataset for SO-101 arm performing language-conditioned pick-and-place tasks in simulation, designed for fine-tuning NVIDIA Isaac GR00T N1.7.
Downloads142
Episodes600
Hours0.959
Why This Matters for Physical AI
This dataset demonstrates vision-language-action learning for robotic manipulation, though it is simulation-only and the sim-to-real gap remains unmeasured, limiting its immediate applicability to physical systems.
Technical Profile
- Modalities
- rgbproprioceptionlanguage
- Robot Embodiments
- SO-101
- Action Space
- 6-DoF: shoulder_pan, shoulder_lift, elbow_flex, wrist_flex, wrist_roll, gripper
- Environment
- simulation
- Task Types
- pick_and_placegrasping
- Episodes
- 600
- Total Hours
- 0.959
- Data Format
- LeRobot
- Annotation Types
- language_instructions
- License
- cc-by-4.0
Access
Need custom rgb data?
Claru builds purpose-built datasets for simulation applications with dense human annotations and quality assurance.
Request a Sample Pack