Charmed MLFlow Beta is here. Try it out now!


Canonical’s MLOps portfolio is growing with a new machine learning tool. Charmed MLFlow 2.1 is now available in Beta. MLFlow is a crucial component of the open-source MLOps ecosystem. The project announced it had passed 10 million monthly downloads at the end of 2022. With Charmed MLFlow users benefit from a platform where they can easily manage machine learning models and workflows.
We are looking for data scientists, machine learning engineers and AI enthusiasts to take Charmed MLFlow Beta for a drive and share their feedback with us.

What is MLFlow?

MLFlow is an open-source platform used for managing machine learning workflows.  It has four primary functions that include experiment tracking, model registry, model management and code reproducibility. Like Kubeflow, MLFlow runs on any environment, including hybrid or multi-cloud scenarios,  and on any Kubernetes.

Try Charmed MLFLow Beta today

Are you a machine learning enthusiast interested in trying MLFlow? Charmed MLFlow uses Juju, Canonical’s Operator Framework, to simplify lifecycle management and integrations. It can easily be integrated with other ML tools such as Charmed Kubeflow.  Charmed MLFlow can be:

  • Deployed on your local machine
  • Used for experiment tracking, model registry, project tracking
  • Improved with your feedback

Deploying Charmed MLFlow is easy with our guide on MicroK8s.

Share your feedback

Charmed MLFlow is an open-source project that is growing because of the care, time and feedback that our community gives. This beta release is no exception, so if you have any feedback or questions, please feel free to contact us.

Give us your feedback

Kubeflow vs MLFlow

Kubeflow and MLFlow are both open-source projects built for machine learning initiatives. They can run on any CNCF-compliant Kubernetes and on any cloud. Kubeflow is designed for ML at scale, handling container orchestration as well as machine learning workflows. It improves the reproducibility of experiments and it allows users to perform the entire machine-learning lifecycle within one tool. MLFlow is a lightweight tool, that allows users to perform training using Python-compatible frameworks.

Both Kubeflow and MLFlow have similar capabilities, but the main difference comes from the ease of setting them up and their end-use cases. Maciej Mazur, Principal AI/ML Field Engineer and Kimonas Sotirchos, Kubeflow Working Group lead,  will host a webinar about Kubeflow vs MLFlow. Are you interested?

Join us to learn more about:

  • Open source MLOps
  • Key differences between Kubeflow and MLFlow
  • How to choose between Kubeflow and MLFlow
Register now

Learn more about MLOps

  • [Whitepaper] A guide to MLOps
  • [Webinar] Hyperparameter tuning with MLOps platform
  • [Blog] A guide to model serving