rllab is a framework for developing and evaluating reinforcement learning algorithms, fully compatible with OpenAI Gym. - View it on GitHub