DIME / Dexterous Imitation Made Easy
NYU-robot-learning
A demonstration-collection and imitation-learning system for dexterous hand–arm manipulation.
What you can explore
Explore nearest-neighbor policy learning and the linked controller, camera and teleoperation packages.
Before you use it
The implementation spans multiple repositories and older ROS/PyTorch dependencies. Set up the controllers and capture stack first; do not treat the model package as a complete robot application.
Original contributors
Sridhar Pandian Arunachalam, Sneha Silwal, Ben Evans and Lerrel Pinto
License / access: Reuse terms not established in the reviewed material; confirm with the original authors
Editorial reference · checked 2026-09-22