deep-chess-playground

Where deep learning meets chess.

This repository aims to implement techniques for neural chess engines, providing an in-depth look at the practical application of AI in chess game.

Table of contents

1. Play chessbots on lichess
2. Run chessbots locally
   • Setup
   • Play
3. Train your own chessbots
   • Data loading
   • Data processing
   • Training

Play chessbots on lichess

COMING SOON

Run chessbots locally

Setup

It is recommended to use Anaconda for managing your Python environment, especially if you plan to use GPU acceleration. However, we also provide instructions for standard Python with pip.

Option 1: Using Anaconda (recommended)

1. Download and install Anaconda from the official website: https://www.anaconda.com/products/distribution.

2. Open Anaconda Prompt and run the following command to create a virtual environment:

   conda create --name deep_chess_playground python=3.10
   

3. Activate the environment:

   conda activate deep_chess_playground
   

4. Install PyTorch. Follow the instructions at PyTorch website. Choose the compute platform (CUDA or CPU) depending on whether you have a GPU or not.

5. Install PyTorch Lightning:

   conda install pytorch-lightning -c conda-forge
   

6. Install deep-chess-playground:

   pip install deep-chess-playground
   

Option 2: Using standard Python and pip

1. Ensure you have Python 3.10 or later installed. You can download it from python.org.

2. Create a virtual environment:

   python -m venv deep_chess_env
   

3. Activate the virtual environment:

   • On Windows:

     deep_chess_env\Scripts\activate
     

   • On macOS and Linux:

     source deep_chess_env/bin/activate
     

4. Install PyTorch. Follow the instructions at PyTorch website. Choose the compute platform (CUDA or CPU) depending on whether you have a GPU or not.

5. Install PyTorch Lightning:

   pip install pytorch-lightning
   

6. Install deep-chess-playground:

   pip install deep-chess-playground
   

Play

COMING SOON

Train your own chessbots

Data loading

If you need a lot of training data, you can use the lichess.org open database which has more than 5 000 000 000 games recorded starting from January 2013!

Training

COMING SOON