How AI describes Humanoid‑Gym

Humanoid-Gym is an easy-to-use reinforcement learning framework, built on Nvidia Isaac Gym, for training locomotion policies for humanoid robots. It emphasizes zero-shot sim-to-real transfer and includes a sim-to-sim verification pipeline from Isaac Gym to MuJoCo to test policy robustness across different physics simulators. It has been demonstrated in real-world pilots on RobotEra's XBot-S and XBot-L humanoids, and its code is available at github.com/roboterax/humanoid-gym, drawing on LeggedGym and rsl_rl resources.