Pre-Release v1.0.2

This version corresponds to the original source code for rsl_rl at the point of publication of "Learning to Walk in Minutes Using Massively Parallel Deep Reinforcement Learning" by Rudin et al.

The release contains an optimized version of PPO implementation suited for use with GPU-accelerated simulators such as Isaac Gym.

This is the version of the code compatible with legged_gym.