Model Card for identifying_pregnancy_clinical_notes

This model is a fine-tuned version of Bio+ClincialBert. It was fine-tuned on an open source Kaggle dataset of medical transcriptions with an aim of classifying clincial notes as discussing or not discussing a patient's pregnancy.

• Repository: Available on GitHub.
• Uses: Intended use is binary classification of English-language clinical notes as discussing or not discussing a patient's pregnancy.
• Limitations: The model was fine-tuned on a small number of clinical notes agumented by limited synthetic data. As a result, it may give inaccurate results at times. It is recommended to consider probability of class assignment when evaluating clinical note classifications.

How to Get Started with the Model

Load the model via the transformers library:

model_path = "elliealbertson/identifying_pregnancy_clinical_notes"
tokenizer = BertTokenizer.from_pretrained(model_path)
model = BertForSequenceClassification.from_pretrained(model_path, num_labels=2)