-
Run
backend/models/identify-symptoms-using-huggingface-transformer.ipynb
, it'll create a fine-tuned-model folder that will contain the model details -
Then simply run the backend
uvicorn main:app --reload
and navigate to Sweeger UI -
Test the
get_symptoms
api to extract symptoms from natural language.
Doctors Image: https://drive.google.com/drive/folders/19drVDiuWjgQXLOfSosXugfibPXGQ0yVH?usp=drive_link
Currently we are using Huggingface Transformer to extract symptoms from text which is supervised learning method.
Next if possible we can try to do some semi-supervised learning for achieveing better accuracy