The Canonical Charmed Kubeflow team is releasing Charmed Kubeflow 1.4 – the state-of-the-art MLOps platform. The new version allows data science teams to securely collaborate on AI/machine learning innovation on any cloud, from concept to production.
Charmed Kubeflow is free to use: The solution can be deployed in any environment without restrictions, firewall, or restricted features.
Data labs and MLOps teams only need to train data scientists and engineers once to work consistently and efficiently on any cloud or on-premise.
Charmed Kubeflow offers a centralized, browser-based MLOps platform that runs on any compatible Kubernetes—providing improved productivity, improved governance, and reduced shadow IT risks, according to the vendor.
The latest release adds several features for advanced model lifecycle management, including upstream Kubeflow 1.4 and support for MLFlow integration.
Data scientists and data engineers can use the MLFlow integration capability to build automatic model drift detection and run a Kubeflow model retraining pipeline. Model skew occurs when model accuracy begins to decline over time due to changes in the direct prediction data set versus the training data set.
Enabling MLFlow on a Kubernetes cluster and integrating it with a Charmed Kubeflow deployment using the unified Juju framework for the launcher is easy, and the MLFlow Juju launcher is available on the CharmHub for immediate deployment.
Charmed Kubeflow 1.4 fully supports multi-user deployment scenarios for all Kubeflow components, including Kubeflow notebooks, pipelines, and experiments. This update simplifies the use of Charmed Kubeflow to improve governance and reduce the incidence of Shadow-IT environments, while helping to combat organizational data leaks.
The Authentication Provider Integration Guide provides more information on setting up multiuser access controls for the Charmed Kubeflow 1.4 MLOps.
The new version is in the stable CharmHub channel now, and can be deployed to any compatible Kubernetes cluster with a single Juju command.
For more information on this news, visit https://canonical.com/.