Every time you run a component whether it contains a machine learning model training, a tensorboard, a hyperparameter optimization, or a workflow, Polyaxon logs that execution as a run and connects related information, such as parameters, artifacts, and upstream or downstream runs.

Overview

All runs in Polyaxon share some characteristics, based on the component/operation specification:

• name, description, and tags
• inputs/outputs

The synthetic acme/demo example shows run metadata, its promoted model versions, and the Lineage > Params tables.

---