Chess AI using Alpha Zero arceticture (supervised learning)

Original source: https://github.com/jackdawkins11/pytorch-alpha-zero

Tried out reinforcement learning, but it was too expensive + too slow.

My article explaining the code

Me 1v1'ing my bot

Training your own chess bot

1. Download this dataset: https://lczero.org/blog/2018/09/a-standard-dataset/

2. Put both train and test in the same folder.

3. Change the folder names in Reformatting.ipynb then run it (this takes a while)

4. Run training.ipynb until the loss can't decrease anymore.

5. Run play.ipynb

Playing with pretrained

1. Run play.ipynb