Downloads · 30 days
23
14% of all-time downloads
medgemma-1.5-4b-it-4bit-autoround
Ashley101179/medgemma-1.5-4b-it-4bit-autoround
medgemma-1.5-4b-it-4bit-autoround is a machine learning model from Ashley101179. Use it for the machine learning task on the model card, and read the license before you ship it in a product. The card lists the license as apache-2.0.
This model is a 4-bit quantized version of google/medgemma-1.5-4b-it, optimized using the Intel AutoRound (Automated Rounding) algorithm.
All-time downloads
161
Public
Parameters
4.3B
3.9 GB on disk
Likes
0
Public
Repo makeup
Click a slice to open those files.
.safetensors3.9 GB · 99%
Parameter types
How the weights are stored.
I323.2B · 75%
Base models
Model card
From the Hugging Face model README
MedGemma-1.5-4B-IT-4bit-AutoRound
🚀 Model Overview
This model is a 4-bit quantized version of google/medgemma-1.5-4b-it, optimized using the Intel AutoRound (Automated Rounding) algorithm.
It is designed for medical reasoning and radiology report generation, offering a significant reduction in memory footprint (from ~10GB down to ~4GB VRAM) while maintaining high inference quality.
🛠 Quantization Details
- Quantization Method: Intel AutoRound
- Bits: 4-bit
- Group Size: 128
- Format: GPTQ
- Backend: compatible with
auto-gptqandoptimum
🩺 Use Cases
- Radiology Report Generation: Describing chest X-rays, CT scans, and other medical imaging findings.
- Medical Consultation: Answering complex medical questions based on clinical guidelines.
- Data Structuring: Extracting key entities from unstructured clinical notes.
💻 How to Use
To run this model, ensure you have transformers, optimum, and auto-gptq installed.
Installation
pip install -U transformers optimum accelerate
pip install auto-gptq --extra-index-url [https://huggingface.github.io/autogptq-index/whl/cu121/](https://huggingface.github.io/autogptq-index/whl/cu121/)
- Type
- gemma2
- License
- apache-2.0
- Created
- Feb 14, 2026
- Updated
- Feb 14, 2026