A set of high-quality implementations of reinforcement learning algorithms created and shared by the OpenAI non-profit AI research company