Preprint
Jul 2026
Extreme-RGMT: Continual Learning of Highly Dynamic Skills for Robust Generalist Humanoid Control
Extreme-RGMT is introduced, a two-stage continual learning framework for robust generalist humanoid control that achieves state-of-the-art generalist whole-body motion-tracking performance, including substantially improved completion of challenging highly dynamic motions.
Yubiao Ma, Han Yu, Kai Guo et al.
· 0 citations