anthonyyazdaniml/gliner-biomed-bi-large-v1.0-medication-regimen-ner

GLiNER-BioMed for medication regimen NER

This model is a fine-tuned version of GLiNER-BioMed-bi-large. This model is designed to extract details about drug administration, dosage, frequency, adverse drug events, and related clinical information, based on the N2C2 2018 Track 2 dataset.

One can find more details about the base GLiNER-BioMed models in the paper GLiNER-BioMed: A Suite of Efficient Models for Open Biomedical Named Entity Recognition. The GLiNER-BioMed code is available at https://github.com/ds4dh/GLiNER-biomed.

Model IDs

• Bi-encoder version (this model): anthonyyazdaniml/gliner-biomed-bi-large-v1.0-medication-regimen-ner
• Uni-encoder version (alternative): anthonyyazdaniml/gliner-biomed-large-v1.0-medication-regimen-ner

Intended use & capabilities

Recognized entity types:

• Drug
• Drug dosage
• Drug strength
• Drug form
• Drug administration route
• Frequency of drug administration
• Treatment duration
• Reason for drug prescription
• Adverse drug event

How to use

First, ensure the gliner library is installed and up-to-date:

pip install gliner -U

Then, you can load and use the model in your Python scripts:

from gliner import GLiNER

model = GLiNER.from_pretrained("anthonyyazdaniml/gliner-biomed-bi-large-v1.0-medication-regimen-ner")

text = """
The patient was prescribed Lisinopril 10 mg tablets to treat her hypertension. 
She was instructed to take one tablet orally once daily for 14 days. 
Shortly after starting the medication, she reported experiencing dizziness.
"""

labels = [
  'Drug', 'Drug administration route',
  'Reason for drug prescription', 'Adverse drug event',
  'Drug dosage', 'Drug strength', 'Frequency of drug administration',
  'Treatment duration', 'Drug form'
]

entities = model.predict_entities(text, labels, threshold=0.5)

for entity in entities:
    print(entity["text"], "=>", entity["label"])

Expected output:

Lisinopril => Drug
10 mg => Drug strength
tablets => Drug form
hypertension => Reason for drug prescription
one => Drug dosage
tablet => Drug form
orally => Drug administration route
once daily => Frequency of drug administration
for 14 days => Treatment duration
dizziness => Adverse drug event

Citation

@misc{yazdani2025glinerbiomedsuiteefficientmodels,
      title={GLiNER-biomed: A Suite of Efficient Models for Open Biomedical Named Entity Recognition},
      author={Anthony Yazdani and Ihor Stepanov and Douglas Teodoro},
      year={2025},
      eprint={2504.00676},
      archivePrefix={arXiv},
      primaryClass={cs.CL},
      url={https://arxiv.org/abs/2504.00676},
}