Robot Reinforcement Training

Train robot policies from video demonstrations.

Assembled an end-to-end video-to-robot learning pipeline (Video2Robot, PromptHMR, MJLab) to train policies from raw video demonstrations using imitation learning techniques