Robot demonstration data
Trajectories collected on real or simulated robots.
- InfinigenOpen source
Procedurally generated photorealistic worlds and indoor scenes.
- Mobile ALOHAOpen source
Whole-body teleoperation and imitation learning on a mobile bimanual robot.
- BlenderProcOpen source
Procedural Blender pipeline for photorealistic training images.
- AgiBot World
One million real humanoid manipulation trajectories across 100+ scenes.
- KubricOpen source
Scalable synthetic video generation with rich ground truth.
- Open X-EmbodimentOpen source
The largest open robot learning dataset and RT-X models.
- RT-1Open source
Robotics Transformer 1 and its 130k-episode dataset of real-world tasks.
- RoboCasa
Large-scale simulated kitchens with generated demonstrations for everyday tasks.
- LAFAN1
Ubisoft's high-quality locomotion mocap used for humanoid retargeting.
- HumanML3DOpen source
Human motions paired with text descriptions for language-to-motion.
- AMASS
Unified human motion capture archive for training humanoid controllers.
- MimicGen
Generate thousands of demonstrations from a handful of human ones.
- DROIDOpen source
A large-scale in-the-wild robot manipulation dataset.
- BridgeData V2Open source
60k+ trajectories of a low-cost WidowX arm across 24 environments.
- RoboNetOpen source
15 million video frames from 7 robot platforms for large-scale visual learning.
- RH20TOpen source
110k contact-rich manipulation episodes with force-torque and multi-view video.
