From the model card
Fields below are copied from the tags and counters on the HuggingFace repository google/medgemma-4b-it at our last fetch. They are set by the uploader, not verified by us; rows with no tag are omitted. How this page is made.
- Publisher (HF namespace)
- Pipeline tag
- image-text-to-text
- Library
- Transformers
- Weight formats
- safetensors
- License tag
other— read the license file in the repo before relying on it- Lineage
-
- base model google/medgemma-4b-pt
- fine-tune of google/medgemma-4b-pt
- Papers cited
- arXiv:2303.15343, arXiv:2507.05201, arXiv:2405.03162, arXiv:2106.14463, arXiv:2412.03555, arXiv:2501.19393, arXiv:2009.13081, arXiv:2102.09542, arXiv:2411.15640, arXiv:2404.05590, arXiv:2501.18362
- Downloads (HF counter at last fetch)
- 1,012,002
- Likes (HF counter at last fetch)
- 1,043
- Model card
- https://huggingface.co/google/medgemma-4b-it
Use cases
- Medical image Q&A for research prototype development
- Radiology report generation assistance in research settings
- Clinical NLP fine-tuning starting point for specific modalities
- Medical AI benchmarking and capability evaluation
Pros
- Multimodal medical specialization in a single 4B model
- Covers four imaging modalities: radiology, dermato, pathology, ophthalmology
- Transformers + TGI compatible for standard deployment
- Apache-like Gemma license for research use
Cons
- Explicitly not approved for clinical use — must not replace physician judgment
- 4B parameters limits reasoning depth on complex clinical cases
- License terms restrict redistribution and derivative model release
- Training data and benchmark comparisons to specialized medical models are limited
Tags
transformerssafetensorsgemma3image-text-to-textmedicalradiologyclinical-reasoningdermatologypathologyophthalmologychest-x-rayconversationalarxiv:2303.15343arxiv:2507.05201arxiv:2405.03162arxiv:2106.14463arxiv:2412.03555arxiv:2501.19393arxiv:2009.13081arxiv:2102.09542