AlphaZero‑inspired chess engine for Apple Silicon, combining an MLX neural network, C++ MCTS UCI engine, and Swift orchestrator for self‑play training, evaluation, and benchmarking.
Stars
1
Forks
0
Watchers
1
Open Issues
1
Overall repository health assessment
No package.json found
This might not be a Node.js project
138
commits
Fix benchmark: add delay for inference server initialization
e7c7555View on GitHubFix policy output format: remove visits count, output move:probability only
89b6b5bView on GitHubFix build.sh to create network directory for setuptools copy operation
1159315View on GitHubFix pybind11 compatibility: update array constructors for newer pybind11 version
a5d423dView on GitHubReduce base MCTS simulations to 400, maintain variable play styles
3373925View on GitHubOptimize AlphaZero training: increase MCTS simulations, adjust temperatures, reduce evaluation timeouts
daee3f3View on GitHubMerge pull request #31 from lukifer23/codex/refactor-mctsengine-for-temperature-handling
57c3e19View on GitHubMerge pull request #30 from lukifer23/codex/refactor-chess-utilities-into-shared-module
be97274View on GitHub