language_table_train_300000_305000_augmented
An augmented subset of the Language Table dataset containing 5,000 episodes with cross-embodiment image augmentations for 8 different robot platforms (Google Robot, Jaco, Kinova3, KUKA iiwa, Panda, Sawyer, UR5e). The dataset includes natural language instructions, end-effector poses, joint positions, and augmented visual observations for policy learning across different robot embodiments.
Why This Matters for Physical AI
This dataset enables training of cross-embodiment robotic policies by providing augmented visual observations and kinematic trajectories across multiple robot platforms with natural language task descriptions, facilitating generalization of learned behaviors across different robot morphologies.
Technical Profile
- Modalities
- rgbproprioceptionlanguage
- Robot Embodiments
- google_robotjacokinova3kuka_iiwapandasawyerur5e
- Action Space
- joint_positions
- Environment
- lab
- Task Types
- manipulation
- Episodes
- 5000
- 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