Medgemma 4b Pt

🚀 MedGemma

MedGemma is a collection of Gemma 3 variants trained for medical text and image comprehension, helping developers accelerate healthcare AI application development.

🚀 Quick Start

If you want to quickly start running the model locally on GPU, follow these steps:

1. First, install the Transformers library. Gemma 3 is supported starting from transformers 4.50.0.

$ pip install -U transformers

2. Here are some example code snippets to help you get started:

Basic Usage

from transformers import pipeline
from PIL import Image
import requests
import torch

pipe = pipeline(
    "image-text-to-text",
    model="google/medgemma-4b-pt",
    torch_dtype=torch.bfloat16,
    device="cuda",
)

# Image attribution: Stillwaterising, CC0, via Wikimedia Commons
image_url = "https://upload.wikimedia.org/wikipedia/commons/c/c8/Chest_Xray_PA_3-8-2010.png"
image = Image.open(requests.get(image_url, headers={"User-Agent": "example"}, stream=True).raw)

output = pipe(
    images=image,
    text="<start_of_image> findings:",
    max_new_tokens=100,
)
print(output[0]["generated_text"])

Advanced Usage

# pip install accelerate
from transformers import AutoProcessor, AutoModelForImageTextToText
from PIL import Image
import requests
import torch

model_id = "google/medgemma-4b-pt"

model = AutoModelForImageTextToText.from_pretrained(
    model_id,
    torch_dtype=torch.bfloat16,
    device_map="auto",
)
processor = AutoProcessor.from_pretrained(model_id)

# Image attribution: Stillwaterising, CC0, via Wikimedia Commons
image_url = "https://upload.wikimedia.org/wikipedia/commons/c/c8/Chest_Xray_PA_3-8-2010.png"
image = Image.open(
    requests.get(image_url, headers={"User-Agent": "example"}, stream=True).raw
).convert("RGB")

prompt = "<start_of_image> findings:"
inputs = processor(
    text=prompt, images=image, return_tensors="pt"
).to(model.device, dtype=torch.bfloat16)

input_len = inputs["input_ids"].shape[-1]

with torch.inference_mode():
    generation = model.generate(**inputs, max_new_tokens=100, do_sample=False)
    generation = generation[0][input_len:]

decoded = processor.decode(generation, skip_special_tokens=True)
print(decoded)

✨ Features

• Medical Focus: MedGemma is a collection of Gemma 3 variants trained for performance on medical text and image comprehension.
• Multiple Variants: It currently comes in two variants: a 4B multimodal version and a 27B text - only version.
• High Performance: Outperforms the base Gemma 3 models across tested multimodal and text - only health benchmarks.

📦 Installation

To use MedGemma, you need to install the Transformers library. Gemma 3 is supported starting from transformers 4.50.0.

$ pip install -U transformers

💻 Usage Examples

Basic Usage

from transformers import pipeline
from PIL import Image
import requests
import torch

pipe = pipeline(
    "image-text-to-text",
    model="google/medgemma-4b-pt",
    torch_dtype=torch.bfloat16,
    device="cuda",
)

# Image attribution: Stillwaterising, CC0, via Wikimedia Commons
image_url = "https://upload.wikimedia.org/wikipedia/commons/c/c8/Chest_Xray_PA_3-8-2010.png"
image = Image.open(requests.get(image_url, headers={"User-Agent": "example"}, stream=True).raw)

output = pipe(
    images=image,
    text="<start_of_image> findings:",
    max_new_tokens=100,
)
print(output[0]["generated_text"])

Advanced Usage

# pip install accelerate
from transformers import AutoProcessor, AutoModelForImageTextToText
from PIL import Image
import requests
import torch

model_id = "google/medgemma-4b-pt"

model = AutoModelForImageTextToText.from_pretrained(
    model_id,
    torch_dtype=torch.bfloat16,
    device_map="auto",
)
processor = AutoProcessor.from_pretrained(model_id)

# Image attribution: Stillwaterising, CC0, via Wikimedia Commons
image_url = "https://upload.wikimedia.org/wikipedia/commons/c/c8/Chest_Xray_PA_3-8-2010.png"
image = Image.open(
    requests.get(image_url, headers={"User-Agent": "example"}, stream=True).raw
).convert("RGB")

prompt = "<start_of_image> findings:"
inputs = processor(
    text=prompt, images=image, return_tensors="pt"
).to(model.device, dtype=torch.bfloat16)

input_len = inputs["input_ids"].shape[-1]

with torch.inference_mode():
    generation = model.generate(**inputs, max_new_tokens=100, do_sample=False)
    generation = generation[0][input_len:]

decoded = processor.decode(generation, skip_special_tokens=True)
print(decoded)

📚 Documentation

Model information

MedGemma is a collection of Gemma 3 variants trained for medical text and image comprehension. Developers can use it to accelerate building healthcare - based AI applications. It currently has two variants: a 4B multimodal version and a 27B text - only version.

Model architecture overview

The MedGemma model is built based on Gemma 3 and uses the same decoder - only transformer architecture as Gemma 3.

Technical specifications

Inputs and outputs

Input:

• Text string, such as a question or prompt
• Images, normalized to 896 x 896 resolution and encoded to 256 tokens each
• Total input length of 128K tokens

Output:

• Generated text in response to the input, such as an answer to a question, analysis of image content, or a summary of a document
• Total output length of 8192 tokens

Performance and validation

MedGemma was evaluated across a range of different multimodal classification, report generation, visual question answering, and text - based tasks.

Imaging evaluations

Chest X - ray report generation

Text evaluations

Ethics and safety evaluation

Evaluation approach

Our evaluation methods include structured evaluations and internal red - teaming testing of relevant content policies. These models were evaluated against categories relevant to ethics and safety, including child safety, content safety, representational harms, and general medical harms.

🔧 Technical Details

Model information

MedGemma 4B utilizes a SigLIP image encoder pre - trained on various de - identified medical data. Its LLM component is trained on diverse medical data. MedGemma 4B is available in pre - trained and instruction - tuned versions. MedGemma 27B is trained exclusively on medical text and optimized for inference - time computation.

Citation

A technical report is coming soon. In the meantime, if you publish using this model, please cite the Hugging Face model page:

@misc{medgemma-hf,
    author = {Google},
    title = {MedGemma Hugging Face}
    howpublished = {\url{https://huggingface.co/collections/google/medgemma-release-680aade845f90bec6a3f60c4}},
    year = {2025},
    note = {Accessed: [Insert Date Accessed, e.g., 2025-05-20]}
}

📄 License

The use of MedGemma is governed by the Health AI Developer Foundations terms of use.

⚠️ Important Note

To access MedGemma on Hugging Face, you're required to review and agree to Health AI Developer Foundation's terms of use. To do this, please ensure you're logged in to Hugging Face and click below. Requests are processed immediately.

💡 Usage Tip

For all MedGemma 27B results, test - time scaling is used to improve performance. If you want to use the model at scale, we recommend that you create a production version using Model Garden.