In MLflow, is there a way group metrics & params into "sections" like the system section?

MLFlow has Model metrics and System Metrics tabs (see attached picutres).

I would like to group metrics by loss and evaluation or example. Is it possible to achieve something similar?

I've tried different naming styles like "group/metric_name" (system metrics are named like that) but they all end up in the Model Metrics section.

I also tried reading through the code to check what could differentiate but there doesn't seem to be anything related to sections in the mlflow python api. Perhaps it's just the way the UI displays it?

As I understand it, there is currently no way to group your params and/or metrics so that it would create another section for the specific run.

The "system metrics" section is separated as the metrics appearing there are logged automatically if you enable the system metric logging as described here: https://mlflow.org/docs/latest/system-metrics/index.html and here: https://mlflow.org/docs/latest/python_api/mlflow.html?highlight=log_metric#mlflow.enable_system_metrics_logging

What you can do

• You can create a subsection with charts in the experiments tab where your runs are listed as here: image example subsection charts .

• See if there is another issue/enhancement idea on the mlflow github, for example this one seems to be related: https://github.com/mlflow/mlflow/issues/9805