GitHub Actions for Machine Learning: Train, Test and Deploy Your ML Model on AWS EC2.
This article is a comprehensive overview of using Github actions to deploy your model into production.
Introduction
Updating a Machine Learning model might require a lot of steps: get and preprocess new data, train your model, and deploy it into production. Those tasks are most of the time repetitive. Wouldn’t it be great to have a tool that can automate those steps? Such tool is Github Actions.
This article starts by providing a good understanding of Github Actions and its benefits. Then, it covers how one can link a Github repository to DagsHub in order to take advantage of tools such as MLFlow, DVC, etc. Finally, it provides a step-by-step approach to creating a complete pipeline from data acquisition, and processing to model deployment on the Amazon EC2 instance and monitoring with MLflow.
What are Github Actions?
Github Actions is a built-in Github tool, used to automate the building, testing, and deploying process of software.
However, Data Scientists and Machine Learning Engineers can use it to automate the entire Data Science workflow, such as :
