mulligan2026MIT
real-square-d2-c00-teleop-baseline
A teleop baseline dataset for real-world square manipulation task (D2), part of the Mulligan robotics benchmark collection. Contains 100 episodes of training data collected from multiple camera viewpoints.
Downloads84
Episodes100
Why This Matters for Physical AI
This dataset provides real-world robot manipulation demonstrations with multi-view observations that enable training and evaluation of embodied AI policies for precise object manipulation tasks.
Technical Profile
- Modalities
- rgb
- Environment
- lab
- Task Types
- manipulation
- Episodes
- 100
- Data Format
- parquet
- Annotation Types
- reward_labels
- 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