PyTorch implementation of BCQ for "Off-Policy Deep Reinforcement Learning without Exploration" - View it on GitHub