This is a PyTorch implementation/tutorial of Deep Q Networks (DQN) from paper Playing Atari with Deep Reinforcement Learning. This includes dueling network architecture, a prioritized replay buffer and double-Q-network training.