A Multi-threaded Implementation of AlphaZero
I spent two days to train a model on GTX TITAN X (similar to GTX1070) see test/models/best_checkpoint
Build Env: ubuntu 20.04 swig4 python3.8 cuda11.7 pytorch1.31.1 libtorch1.31.1
Download packages: LibTorch 1.11 PyTorch 1.11