dl-project

Teaching a Deep Neural Network to solve simple chess endgames solely with Reinforcement Learning.

Dependencies

pip install -r requirements.txt

# macOS local development (included in Google Colaboratory)
pip install --upgrade https://storage.googleapis.com/tensorflow/mac/cpu/tensorflow-0.12.0-py3-none-any.whl

Download Syzygy endgame tablebases

./download_engame_tablebases.sh

Generate self-play data

mkdir -p trainingdata/
python selfplay.py trainingdata/batch1.txt 100 models/v1.h5 models/v2.h5 0.0 [random_seed]