This repository contains the source code for an article about machine learning in production. In the article I explain how a jupyter notebook can be put into production by refactoring preprocessing and inference and using fastapi and onnx runtime. For the preprocessing and inference scikit-learn pipelines are used that can be exported as onnx.
docker build -t loan_image .
docker run --name loan_app -p 80:80 loan_image