InternW0-Δ jointly learns visual dynamics and robot actions
The InternW0-Δ announcement claims the model outperforms prior methods and says its code, weights and data are open-source.
TLDR
The announcement says InternW0-Δ learns visual dynamics and robot actions together and claims it outperforms prior methods. It describes 20K-plus hours of mixed robot, UMI and egocentric data as the largest of its kind. A follow-up links to the project page, repository and paper.
Combined views
7.9K
2 Sources, first seen 9h ago
likes
