ankitjakhar's picture
MedGemma Quantization & Deployment Pipeline
08fe176 verified
|
Raw
History Blame Contribute Delete
421 Bytes
---
license: apache-2.0
base_model:
- google/medgemma-1.5-4b-it
---
# MedGemma 1.5 4B INT4 Benchmark Project
This repository demonstrates:
- NF4 quantization
- Memory optimization
- Benchmark comparison
- GGUF compatibility
- LiteRT conversion
## Results
| Metric | FP16 | INT4 |
|----------|----------|----------|
| VRAM | ~8GB | ~3GB |
| Quality | Baseline | Near-identical |
## Disclaimer
Research purposes only.