hongxiaoymit
RoboTwin 2.0 RGBD Dataset
An RGBD robotic manipulation dataset regenerated from RoboTwin 2.0, augmented with depth camera data for training imitation learning policies such as PI-0. Includes both fixed-scene (clean) and randomized versions covering 50 manipulation tasks.
Downloads13
Episodes550 total (50 clean + 500 randomized)
Why This Matters for Physical AI
This dataset provides high-quality RGBD demonstrations for training imitation learning policies, enabling robot learning from multi-view visual observations with depth information for complex manipulation tasks.
Technical Profile
- Modalities
- rgbdepth
- Robot Embodiments
- dual-arm manipulator
- Action Space
- joint_positions
- Environment
- simulation
- Task Types
- manipulationgraspingpick_and_place
- Episodes
- 550 total (50 clean + 500 randomized)
- Data Format
- LeRobot
- Annotation Types
- language_instructions
- License
- mit
Access
Need custom rgb data?
Claru builds purpose-built datasets for simulation applications with dense human annotations and quality assurance.
Request a Sample Pack