fast-ml-deploy

Example repo of machine learning model deployment with FastAPI and Docker

This is a minimalistic build, consider to use user auth in a production environment

An article explaining the content of this repo can be found in this medium link

Demo instructions

1. Build the docker image

docker build -t iris-ml-build .

2. Run the container

docker run -d -p 80:80 --name iris-api iris-ml-build 

3. Run Pytest with coverage

docker exec -it iris-api pytest --ignore=tests/ --cov=app tests/ --cov-config=.coveragerc

4. Go to localhost

http://127.0.0.1/docs

5. Try out the post /predict method

curl -X POST "http://127.0.0.1/v1/iris/predict" -H\
 "accept: application/json"\
 -H "Content-Type: application/json"\
 -d "{\"data\":[[4.8,3,1.4,0.3],[2,1,3.2,1.1]]}"