mulligan2026MIT
real-routing-d2-legacy-r06-eval-b01
An evaluation dataset of 150 episodes for the real-routing-d2 task, part of the Mulligan robotics benchmark collection. Contains multi-camera observations and trajectory data in parquet format.
Downloads134
Episodes150
Why This Matters for Physical AI
This dataset provides multi-camera real-world robot manipulation trajectories for evaluating routing policies, supporting development and benchmarking of embodied AI systems in controlled task scenarios.
Technical Profile
- Modalities
- rgb
- Environment
- lab
- Task Types
- routing
- Episodes
- 150
- Data Format
- parquet
- Annotation Types
- language_instructionsreward_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