Please see the following blog post for more information on this project https://blog.codecentric.de/en/2018/01/gamma-tictactoe-neural-network-machine-learning-simple-game/
Just clone the repository any run mvn clean compile package
in the gamma-engine directory to compile the project.
In the gamm-engine directory just execute java -jar target/gamma-engine-0.0.1-SNAPSHOT.jar
.
The default configuration can be found from the application.properties in the resources-folder.
// Defines the amount of training runs executed before starting to play, default is 150. training.runs = 150
// Configuration related to selfplay selfplay.enabled = true selfplay.matches = 50 selfplay.games = 10000
learning.stage = 1
humanplay.enabled = true