Show HN: Gemma 4 Multimodal Fine-Tuner for Apple Silicon

Category: ai-ml

Tags: multimodal-ai, fine-tuning, apple-silicon, lora, machine-learning

Score: 6.8/10 (Innovation: 7, Technical: 6, Documentation: 8, Utility: 6)

A multimodal fine-tuning toolkit for Google's Gemma models (text, image, and audio) specifically optimized for Apple Silicon Macs. It's interesting because it uniquely supports audio+text fine-tuning on MPS and can stream training data from cloud storage like GCS/BigQuery, enabling local training on datasets larger than local storage.

Target audience: ai-engineers, ml-researchers, mac-developers

Repository: https://github.com/mattmireles/gemma-tuner-multimodal · Python · MIT · 1147 stars

View on Hacker News