New post: Training an AlphaZero-Style AI for a New Board Game
It goes over everything, architecture, training, deployment. Should be a good reference for anyone wanting to implement their own.
It's a chunky post! A lot of work behind this one.
Link below.
