Track2Act: Predicting Point Tracks from Internet Videos Enables Generalizable Robot Manipulation

Authors:

Abstract
No abstract available