QuaRL is an open-source framework for systematically studying the effect of applying quantization to reinforcement learning algorithms.