ROCm/FastFlowLM v0.9.8

🩺 FastFlowLM v0.9.8 — MedGemma:4B Multimodal Support on AMD Ryzen™ AI

on GitHub

FastFlowLM v0.9.8 brings official support for MedGemma:4B, enabling advanced multimodal (image + text) understanding directly on-device.

🧠 New Model: medgemma:4b

MedGemma is Google’s VLM with image understanding across radiology, pathology, dermatology, and more.

• Run fully offline on your AMD Ryzen™ AI NPU
• Supports multimodal inference using /input <image> prompt in CLI or "images" field in /api/chat

📺 Demo Video:
🎥 MedGemma:4B Running on AMD Ryzen AI NPU (YouTube)

📚 More Info:

• Model page: https://deepmind.google/models/gemma/medgemma/
• Official paper (pg.12–13): https://arxiv.org/abs/2507.05201

⚠️ Disclaimer

This tool (MedGemma + FastFlowLM) is not a diagnostic or clinical tool.
Always consult a licensed medical professional for healthcare decisions.

🔐 Why It Matters

• Privacy First — There is nothing more personal than your health!
• Powered by NPU — Leverages AMD Ryzen™ AI NPU for fast, low-power inference.
• Healthcare Applications — A concrete example of how local LLMs + NPUs enable privacy-preserving, research-driven healthcare workflows.

This release marks a significant leap for on-device multimodal healthcare model deployment and local AI capabilities.