Donghyun1228
LIBERO dense contact corrected yaw 20 degrees
A dataset of 40 task episodes with 346,229 frames from LIBERO tasks in simulation with yaw-20 degree environment compensation. Source actions are replayed deterministically and simulator states are transferred to the compensated environment before rendering.
Downloads844
Episodes40
Why This Matters for Physical AI
This dataset enables study of viewpoint-invariant robotic manipulation by providing controlled environment variations that test generalization of learned policies across different spatial configurations.
Technical Profile
- Modalities
- rgbproprioception
- Robot Embodiments
- Franka Panda
- Action Space
- end_effector_delta
- Environment
- simulation
- Task Types
- manipulation
- Episodes
- 40
- Data Format
- parquet
- Annotation Types
- action_labels
Access
Need custom rgb data?
Claru builds purpose-built datasets for simulation applications with dense human annotations and quality assurance.
Request a Sample Pack