Trying out some different algorithms for chess agents: Minimax and Monte Carlo tree search
Works with Python >= 3.7
Using flask-socketio for graphical play
The game can be run in console mode. This is mainly used for debugging, and for following the steps performed by the agents.
$ python runconsole.py
The game can also be played in the browsed, either manually, by algorithms or a combination.
$ python game.py
If you want to handle one (or both!) players manually, find the lines in the top of runconsole/game that inits a player and set the first parameter to True.
$ pip install -r requirements.txt
Add an agent based on deep neural networks. Download moves, add a fitness function and start training.
Chess moves for download (PGN format)