pepijn223
MolmoAct2-BimanualYAM · generated human demonstration videos
Synthetic human first-person demonstration videos generated from robot episodes of the MolmoAct2-BimanualYAM dataset, designed for Zero-WAM in-context training with 100 episodes across 34 tasks.
Downloads45
Episodes100
Likes1
Why This Matters for Physical AI
This dataset enables training of vision-language models for robotic manipulation by providing synthetic human demonstrations that bridge visual understanding and motor control for bimanual tasks.
Technical Profile
- Modalities
- rgblanguage
- Robot Embodiments
- bimanual manipulator
- Environment
- lab
- Task Types
- manipulation
- Episodes
- 100
- Data Format
- LeRobot
- Annotation Types
- language_instructionsaction_labelssegmentation
Access
Need custom rgb data?
Claru builds purpose-built datasets for lab applications with dense human annotations and quality assurance.
Request a Sample Pack