oxe-auge2025cc-by-4.0
language_table_train_61000_62000_augmented
An augmented subset of the Language Table dataset containing 1,000 episodes with cross-embodiment image augmentations for 8 robot platforms (Google Robot, Jaco, Kinova3, KUKA iiwa, Franka Panda, Sawyer, UR5e) and natural language instructions for task specification.
Downloads77
Episodes1000
Why This Matters for Physical AI
This dataset enables cross-embodiment policy learning by providing synchronized multi-robot view augmentations and language instructions, facilitating the training of generalizable manipulation policies across diverse robot morphologies.
Technical Profile
- Modalities
- rgbproprioceptionlanguage
- Robot Embodiments
- Google RobotJacoKinova3KUKA iiwaFranka PandaSawyerUR5e
- Action Space
- joint_positions
- Environment
- lab
- Task Types
- manipulationtable_manipulation
- Episodes
- 1000
- Data Format
- parquet
- Annotation Types
- language_instructions
- 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