cml

Why we included this project

CML is a command-line tool that hooks model training and evaluation into the CI systems you already run, whether that is GitHub Actions, GitLab CI, or Bitbucket Pipelines. On every pull request it trains the model, checks it against the branch, and posts a Markdown report with metrics and plots right where code review happens, so the numbers are visible before anyone merges. It also covers the awkward parts of MLOps that teams tend to discover late, like provisioning ephemeral cloud runners on AWS, Azure, or GCP for heavier workloads, comparing results against the main branch via DVC, and surfacing TensorBoard links straight from CI. If you want reproducible experiment tracking without standing up a separate ML platform, this is a low-friction way to get there.