Donghyun1228
LIBERO shifted-view rollouts merged for the policy loss
Re-packed LeRobot v2.1 dataset of LIBERO rollouts organized by camera view for policy imitation learning, merging successful and pooled episodes across four suites with consistent episode and frame indexing.
Downloads34
Episodes2035
Why This Matters for Physical AI
Provides a large-scale imitation learning dataset with multi-view observations and action labels for training generalizable manipulation policies across diverse simulated tasks.
Technical Profile
- Modalities
- rgbproprioception
- Action Space
- joint_positions
- Environment
- simulation
- Task Types
- manipulation
- Episodes
- 2035
- Data Format
- LeRobot
- Annotation Types
- language_instructionstask_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