Usage

Clone project

git clone https://github.com/swasun/Yet-Another-AlphaZero
cd Yet-Another-AlphaZero/yaaz/src

You can run the main that runs iteratively self_play, optimisation and evaluator modules:

python3 main.py

Or either run each module separately:

Run the self play module to generate new entries in the dataset.

python3 self_play.py

Run the optimisation module to train the network using the dataset.

python3 optimisation.py

Run the evaluator module to evaluate the best module.

python3 evaluator.py

Dependencies

python3
numpy
keras using a GPU backend (for example: tensorflow-gpu)
python-chess (https://pypi.org/project/python-chess/)
A GPU compatible with CUDA

Cross-plateform

Successfully tested on the following OS (64 bits):

Ubuntu 18.04
Windows 10

References

[1] Silver, D., Schrittwieser, J., Simonyan, K., Antonoglou, I., Huang, A., Guez, A., Hubert, T., Baker, L., Lai, M., Bolton, A., Chen, Y., Lillicrap, T., Hui, F., Sifre, L., van den Driessche, G., Graepel, T. & Hassabis, D. (2017). Mastering the game of Go without human knowledge. Nature, 550, 354--.

[2] Silver, David, Huang, Aja, Maddison, Chris J., Guez, Arthur, Sifre, Laurent, van den Driessche, George, Schrittwieser, Julian, Antonoglou, Ioannis, Panneershelvam, Veda, Lanctot, Marc, Dieleman, Sander, Grewe, Dominik, Nham, John, Kalchbrenner, Nal, Sutskever, Ilya, Lillicrap, Timothy, Leach, Madeleine, Ka-00vukcuoglu, Koray, Graepel, Thore and Hassabis, Demis. "Mastering the Game of Go with Deep Neural Networks and Tree Search." Nature 529 , no. 7587 (2016): 484--489.

[3] Silver, David, Thomas Hubert, Julian Schrittwieser, Ioannis Antonoglou, Matthew Lai, Arthur Guez, Marc Lanctot, Laurent Sifre, Dharshan Kumaran, Thore Graepel, Timothy P. Lillicrap, Karen Simonyan and Demis Hassabis. “Mastering Chess and Shogi by Self-Play with a General Reinforcement Learning Algorithm.” CoRR abs/1712.01815 (2017): n. pag.

[4] C. B. Browne et al., "A Survey of Monte Carlo Tree Search Methods," in IEEE Transactions on Computational Intelligence and AI in Games, vol. 4, no. 1, pp. 1-43, March 2012. doi: 10.1109/TCIAIG.2012.2186810 [5] A general reinforcement learning algorithm that masters chess, shogi, and Go through self-play, D Silver, T Hubert, J Schrittwieser, I Antonoglou, M Lai, A Guez, M Lanctot, L Sifre, D Kumaran, T Graepel, T Lillicrap, K Simonyan, D Hassabis, December 20181