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.
python ai mcts neuralnetwork reinforcementlearning chessai portfolioproject alphazero-inspired montecarlotreesearch
-
Updated
Aug 28, 2026 - Python