Japoivvb/MedGemma-Symptoms

chore: sdk version

chore: provider

feat: use HF inference

fix: cpu?

debug: Move inputs to model device

feat: Update to multimodal MedGemma model

chore:google/medgemma-27b-text-it

debug: remove loading model on buildtime

debug: force CPU

debug: CPU rather

debug: offloading

debug: low mem

debug: faster loading?

chore: update deps

chore: device optimization

chore: auth and logging

fix: use google/medgemma-4b-it

init0

chore: update readme

initial commit