aHiroakiIshikawa2026
Object prompts switched at the d1 lifted-interval midpoint
A dataset of 12 episodes with 5,531 frames (30 fps) featuring dual RGB cameras and 7D robot actions/state, where frame-level task prompts automatically switch from cube targets to coaster targets at the midpoint of detected object lift intervals using a vision language model.
Downloads123
Episodes12
Hours0.0512
Technical Profile
- Modalities
- rgbproprioception
- Robot Embodiments
- 6-DOF manipulator
- Action Space
- 7-DoF: shoulder_pan.pos, shoulder_lift.pos, elbow_flex.pos, wrist_flex.pos, wrist_yaw.pos, wrist_roll.pos, gripper.pos
- Environment
- lab
- Task Types
- manipulationgraspingobject_manipulation
- Episodes
- 12
- Total Hours
- 0.0512
- Data Format
- LeRobot v3.0
Access
Need custom rgb data?
Claru builds purpose-built datasets for lab applications with dense human annotations and quality assurance.
Request a Sample Pack