Deploy ML Model with FastAPI

This project is a part of the Udacity Machine Learning DevOps Engineer Nanodegree.

The key objectives of this project is to apply best practices when it comes to deploying machine learning models.

1. Quality control using pytest.
2. CI/CD using GitHub.
3. Model API deployment using FastAPI.
4. Documentation in the style of Model Cards.

Requirements

Use Python 3.8+.

To install required packages, use the shell command:

pip install -r requirements.txt

Quality control using pytest

I used pytest to perform unittests. Run the tests by using the shell command:

pytest

The above two are automatically run on "push" to GitHub.

Deploying the API locally using FastAPI

Spint up a local API using the shell command:

python -m main

App deployed on Heroku

This app is deployed on Heroku at https://udacitymldevops.herokuapp.com/.

The endpoint can be tested by the command:

python -m heroku_api_request