What is Continuous Machine Learning (CML)?
Continuous Machine Learning (CML) is an open-source library for implementing continuous integration & delivery (CI/CD) in machine learning projects.
It can enable programmatic access to the Machine Learning back-end by creating automated pipelines that simplify model experimentation and deployment while introducing continuous model performance and retraining
This practically means we automate the repetitive process that drives production of these Machine Learning models.
Principles
- GitFlow for data science : We can use GitHub or GitLab to manage ML experiments, track who trained ML models or modified data and when.
- Auto generate reports : We can auto-generate reports with metrics and plots for each Git Pull Request. It helps to make informed decisions for the team.
- No additional services : We can build our own ML platform using just GitHub or GitLab and any cloud provider of our choice.
Advantages
- Changes to our code is connected to fast feedback, to how the code changes have affected our project.
- Since continuous integration gives us feedback instantaneously after our choice, we can make informed decisions to improve the efficiency of the ML model.
- It's a fairly easy process , watch this tutorial to understand how we can set up CML using GitHub actions