← The log

drumih/turbo-fieldfare

★ 3,514 stars

Swift

Gemma 4 26B-A4B inference in ~2 GB of RAM on any M-series MacBook

View on GitHub

18 Aug 2026 · 0:24
TurboFieldfare - Run a 26B AI Model in 2GB of RAM on Any Mac