Jay van Zyl @ ecosystem.Ai

Jay van Zyl @ ecosystem.Ai

Need help training a PPO NN to learn how to play my deckbuilding game

Hey, I have a roguelike deckbuilding game I want to train an agent to play using pure unsupervised RL; I chose PPO as I understand (to my amateur knowledge) that is the most fitting algorithm. I have a very large categorical space that I have to send i…