This service implements the famous TD-Gammon learning system from the Researcher Terald Gesauro.
To gain a deeper understanding check out these sources:
[Wikipedia] (https://en.wikipedia.org/wiki/TD-Gammon)
[Google Scholar] (https://scholar.google.de/scholar?hl=de&as_sdt=0%2C5&q=td+gammon+1992&btnG=&oq=TD-Gammon)
[Official Publication] (https://www.csd.uwo.ca/~xling/cs346a/extra/tdgammon.pdf)
Training the Model:
Either define a proper main method, or just call a function in a module
- cd backgammon_service
- python -m src.<moduleNameHere>
- you might even find a better way....
the model will the be serialialized in the models folder for deploymemt / reuse
Running the Tests:
pip install requirements.txt
pytest test_main.py
Running the Service:
cd backgammon_service
docker run -p 8001:8081 backgammon_service
Contributing:
- In know the project is not perfect (and entirely correct) but colloboration is highly appreciated.