google/medgemma-4b-it

Note

This is a remotely-sourced model from the medgemma plugin, maintained by the community. It is not part of FiftyOne core and may have special installation requirements. Please review the plugin documentation and license before use.

MedGemma is a collection of Gemma 3 variants that are trained for performance on medical text and image comprehension.

Details

• Model name: google/medgemma-4b-it

• Model source: https://huggingface.co/google/medgemma-4b-it

• Model author: Google DeepMind

• Model license: health-ai-developer-foundations (https://developers.google.com/health-ai-developer-foundations/terms)

• Exposes embeddings? no

• Tags: VLM, zero-shot

Requirements

• Packages: huggingface-hub, transformers, torch, torchvision, accelerate, bitsandbytes

• CPU support

  • yes

• GPU support

  • yes

Example usage

import fiftyone as fo
import fiftyone.zoo as foz

foz.register_zoo_model_source("https://github.com/harpreetsahota204/medgemma")

dataset = foz.load_zoo_dataset(
    "coco-2017",
    split="validation",
    dataset_name=fo.get_default_dataset_name(),
    max_samples=50,
    shuffle=True,
)

model = foz.load_zoo_model("google/medgemma-4b-it")

dataset.apply_model(model, label_field="predictions")

session = fo.launch_app(dataset)