https://github.com/saqibali-2k/chessai

https://github.com/saqibali-2k/chessai

Last synced: about 6 hours ago
JSON representation

• Host: GitHub
• URL: https://github.com/saqibali-2k/chessai
• Owner: saqibali-2k
• Created: 2020-07-28T02:55:44.000Z (over 4 years ago)
• Default Branch: master
• Last Pushed: 2022-12-08T11:18:58.000Z (almost 2 years ago)
• Last Synced: 2023-03-04T11:47:28.625Z (over 1 year ago)
• Language: Python
• Size: 43 KB
• Stars: 0
• Watchers: 0
• Forks: 0
• Open Issues: 8
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
# ChessAI

Zero Knowledge Chess AI based off alpha-zero. Uses a neural network and Monte Carlo Tree Search to reduce search depth.

Three main training phases for neural network:

1. Gather data through self-play

2. Train new network

3. Evaluate whether new network is better, replace with the old one if so.

To train:

1. In terminal/cmd, navigate to folder in which you want to clone repo 

2. Clone Repo

3. Run `pip3 install -r requirements.txt`

4. Run `training_pipline.py`

By default, one core on your PC is left free. Change this in training_pipline.py by changing constant CPU_COUNT