Cool project, can we submit models without writing an accompanying paper to go with it?
Though I have to say, using RL to mean "Reinforcement Learning" but also talk about "Roguelike" games (also abbreviated "RL" most of the time) is very confusing.
Though I have to say, using RL to mean "Reinforcement Learning" but also talk about "Roguelike" games (also abbreviated "RL" most of the time) is very confusing.
I had never considered the overloading of the term 'RL' in this situation! I think sadly unavoidable + most people using this benchmark come at it from the reinforcement learning side rather than the roguelike crowd.