MLflow

This is a repackaged open source software product wherein additional charges apply for cloudimg support services.

Overview

MLflow is the widely adopted open source platform for managing the end to end machine learning lifecycle. It provides experiment tracking to log parameters, metrics and artifacts, a model registry to version and stage models, and tools to package and deploy them. This image delivers the MLflow tracking server and web UI fully installed and configured as a system service, so a production ready ML platform is running within minutes of launch. The current release available is MLflow 3.13.

Application Stack

MLflow is installed into a dedicated Python virtual environment under /opt/mlflow and run by an unprivileged service account on Python 3.12. The tracking server listens on the loopback address and an nginx reverse proxy fronts it on port 80. A systemd service starts the server on boot and restarts it on failure.

Secure By Default

The UI and REST API are protected by HTTP Basic Authentication. This image generates a fresh administrator password, unique to your instance, on its first boot and writes it to a root only file. The unauthenticated health probe stays open for load balancers; everything else requires the password. No shared or default credentials ship in the image.

Ready To Use

Point your training code at the instance on port 80 with the MLflow client, log experiments and register models, and browse them in the web UI. The backend store and artifact store live on a dedicated, independently resizable storage volume kept separate from the operating system disk. For production scale, repoint the backend store to PostgreSQL and the artifact store to Amazon S3.

cloudimg Support

24/7 technical support by email and chat. Help with MLflow deployment, experiment tracking, the model registry, backend and artifact store configuration, TLS termination and scaling.

Use Cases

Centralised experiment tracking for data science teams. A model registry and staging workflow. A self hosted, in your own VPC MLOps platform for teams with data residency or compliance requirements. Reproducible machine learning pipelines.

All product and company names are trademarks or registered trademarks of their respective holders. Use of them does not imply any affiliation with or endorsement by them.