Ready-made tasks to train and evaluate on.
Reinforcement learning environments for Unitree Go2, H1 and G1 robots.
GPU-accelerated environments for locomotion and manipulation in MJX.
