legged_gym
Active
legged_gymRLreinforcement learninglocomotionETH RSLIsaac GymANYmal
legged_gym is an open-source learning environment developed by Nikita Rudin and Marco Hutter at the Robotic Systems Lab (RSL) ETH Zurich, released alongside the paper "Learning to Walk in Minutes Using Massively Parallel Deep Reinforcement Learning" (CoRL 2021). Built on NVIDIA Isaac Gym, it enables training locomotion policies for legged robots (ANYmal, bipeds) on a single GPU in minutes.