Hannibal52Barcaapache-2.0
Demo Action Space
A bimanual teleoperation dataset derived from the ICL demo dataset, augmented with Cartesian end-effector poses computed via forward kinematics using 6D rotation representation.
Downloads0
Episodes285
Why This Matters for Physical AI
This dataset enables training of imitation learning models on bimanual manipulation tasks with continuous Cartesian pose representations, which is fundamental for learning generalizable manipulation policies.
Technical Profile
- Modalities
- rgbproprioception
- Robot Embodiments
- bimanual manipulator
- Action Space
- end_effector_delta
- Environment
- lab
- Task Types
- manipulation
- Episodes
- 285
- Data Format
- parquet
- Annotation Types
- action_labelstask_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