sidupadhyay/bert-base-uncased_clinical-ner-ONNX

Clinical NER (ONNX)

Available on Hugging Face: sidupadhyay/bert-base-uncased_clinical-ner-ONNX

This is an ONNX export of samrawal/bert-base-uncased_clinical-ner.

All credit for model training and development goes to the original author, samrawal.

Model Details

• Architecture: BertForTokenClassification (bert-base-uncased)
• Task: Named Entity Recognition (NER) on clinical text
• Format: ONNX
• Max sequence length: 512

Labels

Label	Description
B-problem / I-problem	Clinical problem or diagnosis
B-treatment / I-treatment	Treatment or medication
B-test / I-test	Clinical test or procedure
O	Outside any entity

Usage

from transformers import AutoTokenizer
from optimum.onnxruntime import ORTModelForTokenClassification

tokenizer = AutoTokenizer.from_pretrained("sidupadhyay/bert-base-uncased_clinical-ner-ONNX")
model = ORTModelForTokenClassification.from_pretrained("sidupadhyay/bert-base-uncased_clinical-ner-ONNX")

inputs = tokenizer("The patient was prescribed aspirin for chest pain.", return_tensors="pt")
outputs = model(**inputs)

Original Model

For more details on training data, evaluation, and intended use, see the original model card: samrawal/bert-base-uncased_clinical-ner.