You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
An AlphaZero-style chess engine in Python: PUCT tree search guided by a policy and value network that learns entirely from self-play. The whole pipeline is here, self-play to replay data to training loop to checkpoints, and it plays from the command line.