Jinxy Chess AI

A powerful neural network chess engine with advanced AI capabilities.

Explore Jinxy Chess AI Docs »

Report bug · Open a discussion · Discord · Blog

NOTE

this repository is only to read code purposes , after cloning u wont be able to play cuz there are some functionalities that i removed for the purpose of more development

Overview

Welcome to the Jinxy Chess AI Project! This project integrates a sophisticated machine learning-based artificial intelligence (AI) with a user-friendly graphical user interface (GUI) built using Pygame. The goal is to provide an engaging and challenging chess-playing experience where users can compete against an AI opponent capable of making intelligent and strategic moves based on a trained neural network model

Introduction

Chess has been a cornerstone in the development of artificial intelligence, serving as a benchmark for testing AI capabilities in strategic thinking and decision-making. The Chess AI Project aims to merge the strategic depth of chess with modern AI techniques, creating an AI opponent that can challenge players of varying skill levels.

This project leverages a neural network trained on extensive chess game data to predict and execute moves. The integration with Pygame ensures that users have a seamless and interactive experience, allowing them to focus on the game without being bogged down by technical complexities.

Whether you're a chess enthusiast eager to test your skills against an AI or a developer interested in the intersection of game development and machine learning, this project offers valuable insights and a robust platform for exploration.

---

Features

• Interactive GUI:

  • Visually appealing chessboard rendered using Pygame.
  • Real-time move highlighting for selected pieces and valid destinations.
  • Smooth animations for piece movements and special actions.

• AI Opponent:

  • Utilizes a neural network model trained on diverse chess game data.
  • Capable of predicting and executing intelligent moves based on the current board state.
  • Adjustable difficulty levels to cater to players of different skill sets.

• Move Validation:

  • Comprehensive validation ensuring adherence to official chess rules.
  • Supports all standard moves, including castling, en passant, and pawn promotion.
  • Prevents illegal moves and provides feedback to the user.

• Game State Management:

  • Maintains accurate game state using FEN (Forsyth-Edwards Notation).
  • Detects game-ending conditions such as checkmate, stalemate, threefold repetition, and the fifty-move rule.
  • Provides options to restart the game or exit upon conclusion.

• Move History:

  • Logs all moves made during the game for reference and analysis.
  • Displays move history within the GUI for easy access.

• Promotion Handling:

  • Prompts users to choose a piece upon pawn promotion.
  • Ensures that promotions are executed smoothly without disrupting gameplay.

• Customization:

  • Ability to change board colors and piece styles.
  • Support for different image sets to personalize the gaming experience.

• Performance Optimization:

  • Efficient neural network architecture for quick move predictions.
  • Utilizes CPU or GPU resources based on availability for enhanced performance.

---

• board.py:

  • Maintains the game state using the Board and Move classes.
  • Handles move generation, validation, and state updates.
  • Manages special moves like castling, en passant, and promotion.

• predict_move.py:

  • Interfaces with the neural network model to predict the best moves based on the current board state.
  • Converts FEN strings to numerical feature vectors suitable for the neural network.

• minimax.py:

  • Implements the Minimax algorithm as a fallback for move suggestions if the neural network does not provide a move.
  • Provides functions to evaluate board states and determine optimal moves.

• train_model.py:

  • Script for training the neural network model using training data from PGN files.
  • Handles data preprocessing, feature engineering, model training, and saving of trained models.

• data/:

  • Directory for storing training PGN (Portable Game Notation) files.
  • Collect a diverse set of chess games to train the AI effectively.

• LICENSE:

  • Contains the licensing information for the project. The current project uses the MIT License.

---

Installation

Setting up the Chess AI Project involves several steps, including installing necessary dependencies, setting up a virtual environment, and ensuring that all required assets are in place.

Prerequisites

Before diving into the installation, ensure that your system meets the following prerequisites:

• Operating System:

  • Windows 10 or higher
  • macOS Catalina or higher
  • Linux (Ubuntu 18.04 or higher recommended)

• Python Version:

  • Python 3.7 or higher
  • Download Python from python.org

• Git:

  • For cloning the repository
  • Download Git from git-scm.com

• Hardware:

  • CPU: Modern processor
  • GPU (Optional): For accelerated neural network training and inference using PyTorch