MLflow

The MLflow app launches MLflow Tracking, a user interface for tracking machine learning runs.

To launch it, select the log directory where you have data to visualize and the resources for the Slurm job. The log directory must be a valid directory with logs for MLflow to function correctly.

Note it is also possible to deploy MLflow as a service on Kubernetes (using LUMI-K Cloud), more details are available in the short tutorial.