MultiClinNER-UniboNLP/medgemma-en-ner-en-disease-3epochs-clean
The MultiClinNER-UniboNLP/medgemma-en-ner-en-disease-3epochs-clean is a 4.3 billion parameter Gemma3 model, developed by MultiClinNER-UniboNLP. This model is fine-tuned for English Named Entity Recognition (NER) specifically for disease entities. It was trained using Unsloth and Huggingface's TRL library, enabling faster fine-tuning. Its primary strength lies in accurately identifying and extracting disease-related information from text.
Loading preview...
Model Overview
MultiClinNER-UniboNLP/medgemma-en-ner-en-disease-3epochs-clean is a specialized 4.3 billion parameter model developed by MultiClinNER-UniboNLP. It is fine-tuned from the unsloth/medgemma-4b-it-bnb-4bit base model, leveraging the Gemma3 architecture. This model was specifically trained for English Named Entity Recognition (NER) with a focus on identifying disease entities within text.
Key Capabilities
- Disease Named Entity Recognition: Excels at identifying and extracting disease names and related medical conditions from English text.
- Efficient Fine-tuning: Benefits from being trained 2x faster using Unsloth and Huggingface's TRL library, indicating an optimized training process.
- Gemma3 Architecture: Built upon the Gemma3 model family, providing a robust foundation for its NER capabilities.
Good For
- Medical Text Analysis: Ideal for applications requiring the extraction of disease information from clinical notes, research papers, or other medical documents.
- Biomedical Information Extraction: Useful in scenarios where precise identification of disease entities is crucial for downstream tasks like knowledge graph construction, clinical decision support, or epidemiological studies.
- Research and Development: Provides a specialized tool for researchers working on medical NLP tasks, particularly those involving disease entity recognition.