mulligan2026MIT
real-marker-d2-c04-dagger-baseline
A robotics dataset from the Mulligan project containing 50 episodes of marker-based manipulation tasks collected with multiple camera views for training robotic policies using DAGGER (Dataset Aggregation) approach.
Downloads280
Episodes50
Hours0.2527
Why This Matters for Physical AI
This dataset enables training of robotic manipulation policies through multi-view camera observations and demonstrates the DAGGER methodology for iterative policy improvement in real-world robotic tasks.
Technical Profile
- Modalities
- rgb
- Robot Embodiments
- franka
- Action Space
- 7-DoF: vel_x, vel_y, vel_z, vel_roll, vel_pitch, vel_yaw, gripper_action
- Environment
- lab
- Task Types
- manipulation
- Episodes
- 50
- Total Hours
- 0.2527
- 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