deluca Performant, differentiable reinforcement learning NOTE: this is pre-alpha software and is undergoing a number of core changes. Updates to follow.