Scalable, open-source, reinforcement learning library focused on modularity and simplicity for easy hacking. It currently includes all scalable versions of PPO and SAC.
