Show HN: Chesslm – Tiny chess model trained on Stockfish

Category: ai-ml

Tags: chess, neural-network, game-playing, alpha-beta, numpy

Score: 5.3/10 (Innovation: 5, Technical: 6, Documentation: 7, Utility: 3)

Chesslm is a tiny neural chess evaluator (100k parameters) that combines a NumPy network with a fixed material evaluator and alpha-beta search to play chess. It includes desktop (Tkinter), web (Python server), and browser (Pyodide) interfaces, plus scripts for training on Stockfish-generated data. The project is interesting for its compact model and educational value in showing how a small neural network can achieve moderate playing strength.

Target audience: AI enthusiasts, hobbyist chess programmers, and developers interested in small neural networks

Repository: https://github.com/skorotkiewicz/chesslm · Python · 2 stars

View on Hacker News