bvanaken / clinical-outcome-prediction

Code for the EACL 2021 Paper: Clinical Outcome Prediction from Admission Notes using Self-Supervised Knowledge Integration
Apache License 2.0
87 stars 28 forks source link

Huggingface model inference (diagnosis prediction) #16

Closed ashishakkumar closed 2 months ago

ashishakkumar commented 2 months ago

@bvanaken Can you please give an example code of how to using the model on HuggingFace you uploaded, to use for the diagnosis prediction. Moreover, may I ask where is the evaluation script located for the diagnosis prediction and the other comparisons (BioBERT, ClinicalBERT,etc) Thank you

bvanaken commented 2 months ago

Hi, For diagnosis prediction, I suggest using the model fine-tuned on the task. You can find it here: https://huggingface.co/DATEXIS/CORe-clinical-diagnosis-prediction. The base CORe model is not fine-tuned on a specific task yet, but pre-trained using clinical outcome prediction.

For evaluating the other models, you can simply use them as base models in the doc_classification.py script and you will get results after fine-tuning them to the task you want to evaluate.

Hope this helps! Betty

ashishakkumar commented 2 months ago

Thank you