naavox2026
simple_grasp_spin
Extended version of the simple_grasp dataset with recovered gripper camera heading (spin) information added to observation states. Contains 370 episodes with 148,916 frames of robotic grasping and manipulation data.
Downloads48
Episodes370
Hours1.3789
Why This Matters for Physical AI
This dataset enables visual servoing research by providing complete camera pose information (gripper heading) necessary for projecting room coordinates into camera frames and training visuomotor control policies.
Technical Profile
- Modalities
- rgbproprioception
- Robot Embodiments
- mobile_manipulator
- Action Space
- 5-DoF: vel_x, vel_y, vel_z, wrist_speed, finger_speed
- Environment
- lab
- Task Types
- graspingmanipulation
- Episodes
- 370
- Total Hours
- 1.3789
- Data Format
- LeRobot
- Annotation Types
- action_labels
Access
Need custom rgb data?
Claru builds purpose-built datasets for lab applications with dense human annotations and quality assurance.
Request a Sample Pack