Ready-made tasks to train and evaluate on.
Reinforcement learning environments for Unitree Go2, H1 and G1 robots.
GPU-parallel RL environments that popularized massively parallel sim-to-real.
