Model Overview

MultiClinNER-UniboNLP/medgemma-en-ner-en-disease-3epochs-COT is a specialized Gemma-3 based language model developed by MultiClinNER-UniboNLP. It has been fine-tuned for English Named Entity Recognition (NER), with a specific focus on identifying disease entities within text.

Key Characteristics

• Architecture: Based on the Gemma-3 model family.
• Specialization: Fine-tuned for English disease Named Entity Recognition (NER).
• Training Efficiency: Training was accelerated using Unsloth and Huggingface's TRL library, resulting in 2x faster finetuning compared to standard methods.
• Origin: Finetuned from the unsloth/medgemma-4b-it-bnb-4bit model.

Use Cases

This model is particularly well-suited for applications requiring precise identification of disease names in English-language medical, clinical, or biomedical texts. Potential use cases include:

• Automated extraction of disease mentions from electronic health records (EHRs).
• Populating medical knowledge bases with disease information.
• Assisting in clinical trial recruitment by identifying relevant patient conditions.
• Supporting epidemiological studies by extracting disease prevalence data from text.