An open API service indexing awesome lists of open source software.

https://github.com/auraluvsu/nyx

Nyx is a blazing fast chess engine written in Go, complete with machine learning in Python using PyTorch
https://github.com/auraluvsu/nyx

ai chess-engine go golang golang-application machinelearning machinelearning-python pytorch

Last synced: 3 months ago
JSON representation

Nyx is a blazing fast chess engine written in Go, complete with machine learning in Python using PyTorch

Awesome Lists containing this project

README

          

Project Nyx


Class level: Omega


Nyx is a blazing fast Go-powered chess engine fused with a PyTorch-based ML Layer
for smart move predictions and advanced evaluation


What is Project Nyx Exactly?


Project Nyx is a hybrid chess engine built for speed, simplicity and efficiency:


  • Go handles the engine core. Move generation, board representation,
    evaluation and legality checking.

  • Python, specifically PyTorch powers the ML model. Its trained to suggest
    moves based on game state, learning from classical and modern playstyles


This engine is designed to be lightweight, extensible, efficient and smart enough
to give you a real challenge, and get smarter by learning from your games.

Features:



  • Engine written in pure Go for its performance and lightweight concurrency model

  • Integrated PyTorch model for position evaluation and optimal move suggestion

  • Self-play + training loop support

  • Load/save game state via FEN/PGN


Prerequisites:



  • Python 3.10+

  • Go 1.20+

  • PyTorch (with CUDA support if available)


How the AI works


The engine sends board states to the ML model via shared memory. The PyTorch
model returns evaluation scores and candidate moves, which the Go engine uses in its
search algorithm.
You can train the model on:

  • Historical PGN datasets (Grandmaster games, Lichess, etc.)

  • Self-play games using reinforcement learning