CML, or “Continuous Machine Learning,” is a library to monitor changes in ML models using continuous integration and delivery (CI/CD) paradigms. Specifically, users can customize their model evaluations, automate comparisons across experiments and versions, and track against changing datasets. CML is built on top of “Gitflow” principles and uses DVC to track data and models. Reports with custom plots and metrics can be registered to auto-generate with every git pull request.