PCOS Prediction API built with FastAPI and deployed on render.

Machine learning model built with Google Colab.

Docs available at /docs.

1. Clone repository

git clone https://github.com/ranmerc/pcos-prediction-backend.git

2. Create Virtual Environment

# Creates Virtual Environment named venv
python -m venv venv

3. Activate venv

source venv/Scripts/activate

4. Install packages

pip install -r requirements.txt

5. Running Uvicorn Server

uvicorn main:app --reload

6. Deactivating Virtual Environment

deactivate

• Generating requirements.txt

pip freeze > requirements.txt

We pickle the trained model and StandardScaler object -

import pickle
pickle.dump(model, open("model.sav", 'wb'))
pickle.dump(standard_scalar, open("sc.sav", 'wb'))

Then we load the model on server and make prediction using it -

loaded_model = pickle.load(open("model.sav", 'rb'))
sc = pickle.load(open("sc.sav", 'rb'))

test_data = []
test_data = sc.transform([test_data])
res = loaded_model.predict(test_data)

• Can not activate a virtualenv in GIT bash mingw32 for Windows on Stack Overflow

• Is it bad to have my virtualenv directory inside my git repository? on Stack Overflow

• Python Tutorial: VENV (Windows) - How to Use Virtual Environments with the Built-In venv Module by Cory Schafer on Youtube

• Deploying FastAPI application to Render by Akash R Chandran

• What and why behind fit_transform() and transform() in scikit-learn! by Towards Data Science