Skip to content

kind-lab/transformer-deid

Repository files navigation

transformer-deid

Fine tune transformer models to deidentify clinical medical data.

Setup

Install dependencies in a conda environment:

conda env create -n transformer_deid --file environment.yml

Data

Data must be in CSV stand-off format: a subfolder (txt/) contains the documents in individual text files with the document identifier as the file stem and .txt as the extension. Another subfolder (ann/) contains a set of CSV files with the annotations with the same document identifier as the file stem and .gs as the extension. The tests/data subfolder contains an example of documents stored in this format.

Training

Models supported:

  • BERT
  • DistilBERT
  • RoBERTa

To run from the repository directory,

python transformer_deid/train.py -m <model_architecture> -i <dataset path> -o <output path> -e <number of epochs>

Options:

  • -m --model_architecture Name of model {bert | distilbert | roberta}.
  • -i --train_path Path to dataset directory.
  • -o --output_path Model save directory.
  • -e --epochs Number of epochs.

Evaluation

For evaluation, see Pyclipse.