Team-RAMENcc-by-4.0
RAMEN G1 IKEA Assembly — RAMEN-Ori rotate_leg_to_tighten
A curated subset of 290 episodes (232,634 frames) from the BitRobot G1 IKEA assembly dataset, focusing on the rotate_leg_to_tighten skill for the Unitree G1 humanoid robot with Dex1-1 gripper. Includes multi-camera RGB/IR imagery with YOLO-OBB object detection overlays and augmented frame variants for training the RAMEN-Ori policy.
Downloads21
Episodes290
Hours2.2
Why This Matters for Physical AI
This dataset provides high-quality multi-modal demonstrations of fine-grained assembly manipulation for humanoid robots, with object detection annotations that enable learning robust visuomotor policies for complex real-world tasks.
Technical Profile
- Modalities
- rgbinfraredproprioception
- Robot Embodiments
- Unitree G1humanoid
- Action Space
- joint_positions
- Environment
- lab
- Task Types
- manipulationassembly
- Episodes
- 290
- Total Hours
- 2.2
- Data Format
- LeRobot
- Annotation Types
- bounding_boxescuration_labelsobject_detection
- License
- cc-by-4.0
Access
Need custom rgb data?
Claru builds purpose-built datasets for lab applications with dense human annotations and quality assurance.
Request a Sample Pack