HyeonseokEapache-2.0
Rq2 Stack 2 Cubes Random 80 10Fps
LeRobot v3.0 dataset of a simulated so101 robot stacking blocks, collected via SCRAPE-IsaacLab pipeline running in Isaac Sim 5.1/IsaacLab 2.3.2 with per-frame natural language skill labels.
Downloads16
Episodes80
Why This Matters for Physical AI
This dataset demonstrates the use of large language models (Gemini) for automatic skill annotation of robotic trajectories, enabling scalable generation of semantically-labeled manipulation data for training embodied AI agents.
Technical Profile
- Modalities
- rgblanguage
- Robot Embodiments
- so101_follower
- Environment
- simulation
- Task Types
- manipulationstacking
- Episodes
- 80
- Data Format
- LeRobot
- Annotation Types
- language_instructionsskill_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