grl-framework This is a generic reinforcement learning framework. At the moment, only finite action spaces are allowed.