GenRobot releases Gen-HumanEgo: 1,800+ hours of egocentric human data for robot learning
AdinaYakup · x · 2026-09-24
GenRobot has released Gen-HumanEgo on Hugging Face, a large-scale dataset for embodied AI and robot learning:
- 1,800+ hours of real-world egocentric human demonstrations
- 10K+ unique tasks across diverse environments, people, and activity styles
- Captured with a unified six-camera DAS-Ego setup with 6 synchronized views
- Structured supervision includes 3D hand keypoints, MANO hand meshes, egocentric depth estimation, and task annotations
Recordings are processed by GenRobot's Data Foundation Model (DFM) to produce complementary training signals for human motion and spatial understanding, targeting VLA, manipulation, and SLAM research. Licensed CC-BY-SA-4.0.
Related event: GenRobot Open-Sources Gen-HumanEgo: 1,800-Hour Egocentric Robot Dataset(2 posts)→
More from Embodied
- Uranus: diffusion-based robot simulator hits 24 FPS with open-ended rollout — D-Robotics · 2026-09-24
- Microsoft Open-Sources LIBERO Robot Learning Dataset With 282K Rows on Hugging Face — victormustar · 2026-09-24
- BitRobot to host 3 robotics grand challenges at IROS 2026 with 50+ research teams — micoolcho · 2026-09-24
- Meta reportedly made a keychain just for talking to AI — prasenx · 2026-09-24
- TJ-FlyingFish: Chinese team builds 1.6kg amphibious drone that flies and dives for 40 minutes — lukas_m_ziegler · 2026-09-24
- Mifeng launches China's answer to Figure's Index: crowdsourced robot data platform with 1M hours — 量子位 · 2026-09-24