NVIDIA introduces GR00T 2 robot world action model at CoRL
Original title: Practice Makes Precise: At CoRL, NVIDIA and Partners Reveal How Robots Rehearse Delicate Work in Simulation
NVIDIA says GR00T 2 hit 71% average success across 50 tasks with one demonstration and 95% with five.
At the Conference on Robot Learning in Austin, NVIDIA introduced open robotics technologies centered on GR00T 2, a world action model built on NVIDIA Cosmos and trained on robot and human data, including 100,000 hours of open-source human egocentric data. The company also highlighted Isaac Lab 3.0, tighter Newton physics integration, the EgoHand model, and a Video to Data workflow. Two NVIDIA papers are among six Best Paper Award finalists. Partners and testers named include Siemens, Foxconn, Nexuni, and MicroAGI.
Unconfirmed: Task-success and leaderboard figures are NVIDIA's own reported results.
- GR00T 2 completed 40 of 50 evaluated tasks at least once without task-specific demonstrations.
- NVIDIA ranks GR00T 2 among the top three on RoboCasa365, RoboLab, RoboDojo, and MolmoSpaces.
- Isaac Lab 3.0 with Newton reached up to 2.5x the training throughput of Isaac Lab 2.3 with PhysX.
- EgoHand had over 40,000 Hugging Face downloads in the two weeks since launch.
- GRAIL and RoboTTT are two of six CoRL Best Paper finalists.
- Vendors
- NVIDIA
- Products
- Isaac LabNewtonNVIDIA Cosmos
From the original article
7 min read at blogs.nvidia.com
Simulation, which has long been the proving ground for robots learning to walk, roll and hop, is now helping researchers teach robots tasks that require a finer touch: aligning gears, connecting cables and assembling small components. The shift from locomotion to manipulation is taking center stage at this week’s Conference on Robot Learning (CoRL) in […]
By Sasa Docca · blogs.nvidia.com
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.