Graph-AlphaZero

About The Project

This project aims to train an alphazero-style engine for Chess. The biggest difference being the policy network; we use a Graph Attention Network

Report
Report (Short version)

Name		Name	Last commit message	Last commit date
Latest commit History 42 Commits
plots		plots
train_human_games		train_human_games
utils		utils
.DS_Store		.DS_Store
README.md		README.md
array_net.py		array_net.py
buffer.py		buffer.py
datamodule.py		datamodule.py
eval.py		eval.py
finetune.py		finetune.py
gui.py		gui.py
hpo.py		hpo.py
mcts.py		mcts.py
network.py		network.py
old_network.py		old_network.py
pgn.txt		pgn.txt
play.py		play.py
plot.ipynb		plot.ipynb
requirements.txt		requirements.txt
stock.py		stock.py
train.py		train.py