Distributed online training of a general multi-task Deep RL Agent - View it on GitHub