biomed-ner-transformer-v3

Overview

This model is a fine-tuned BERT-base-cased transformer specialized for Named Entity Recognition (NER) in biomedical texts. It accurately identifies drugs, diseases, and gene sequences in clinical notes and research abstracts.

Model Architecture

• Base Model: bert-base-cased
• Output Layer: Linear Token Classification head with 7 labels (BIO format).
• Training: Fine-tuned on 100k+ annotated PubMed snippets.

Intended Use

• Automating clinical data extraction.
• Parsing medical journals for knowledge graph construction.
• Assisting in medical coding workflows.

Limitations

• Performance may vary on handwritten clinical notes due to spelling variances.
• It is not a diagnostic tool and should only be used for research/informational purposes.

Example Code

from transformers import pipeline

ner_pipeline = pipeline("ner", model="biomed-ner-transformer-v3", aggregation_strategy="simple")
text = "Patient was prescribed Metformin for Type 2 Diabetes treatment."
print(ner_pipeline(text))