Earl: a framework for scalable reinforcement learning research

This blog post describes Earl, a reinforcement learning (RL) framework that enables scalable distributed training across multiple devices.