fokan
/

MedFusion-AI

Model card Files Files and versions

MedFusion-AI / lite /decoder /README.md

fokan's picture

Add unified MedFusion-AI (Pro+Lite) with pipeline & app

b85ba72 verified 3 months ago

|

history blame contribute delete

203 Bytes

fokan/medgemma-4b-it-int8

INT8 dynamic quantized version of google/medgemma-4b-it

Quantization: Dynamic INT8 on Linear layers (PyTorch)
Ideal for CPU inference
4× smaller than original model