Repository navigation
Lozza 3.0
Lozza goes neural. The hand-crafted evaluation is replaced by a small NNUE integrated from my Cwtch experiment: a white-relative 768 → 75 → 1 network with SReLU activation and float weights, trained with my own Javascript trainer (trainer.js).
Rating
CCRL 40/40: 2890 · CCRL Blitz: 2831
Usage
Use lozza.js with Node.js in any UCI chess GUI (Arena, Cute Chess, etc.), or in a web page/web worker.
Binaries
While Lozza is intended for web use, I can supply binaries on request. Please email me op12no2@gmail.com noting the target platform; thanks.