Zeus-OmniMIT
D3IL Vision + LeRobot / LingBot-VA
A derivative of D3IL demonstrations with two-camera vision coverage across eight task subsets, converted to LeRobot v2.1 format with precomputed LingBot-VA video/text features for imitation learning.
Downloads285
Episodes8,298
Likes1
Why This Matters for Physical AI
This dataset provides large-scale multi-task imitation learning data with vision and language features on a collaborative robot arm, enabling training of generalizable manipulation policies across diverse contact-rich tasks.
Technical Profile
- Modalities
- rgbproprioceptionlanguage
- Robot Embodiments
- Franka Panda
- Action Space
- joint_positions
- Environment
- labsimulation
- Task Types
- manipulationavoidingpushinginsertingaligningsortingstacking
- Episodes
- 8,298
- Data Format
- LeRobot
- Annotation Types
- language_instructions
- License
- MIT
Access
Need custom rgb data?
Claru builds purpose-built datasets for lab applications with dense human annotations and quality assurance.
Request a Sample Pack