Model Lifecycle

AI Observatory allows users to enable important features by integrating with their model lifecycle tracking process. Users can easily enable model lifecycle tracking if they are using MLflow, Metaflow, etc.

Integrating Model Versions

Once model versions are tracked by AI Observatory, the user can enable the following functionality:

• Model comparison: two versions of the same model that have been monitored over the same period of time can be compared in the performance dashboard.
• Model output comparison: two versions of the same model that have been monitored over the same time can be compared along core outputs which they are producing.
• Model version indicators: once a model version tag is updated to a new version, an indicator will be displayed on all time-specific plots in the performance dashboard, output feature pages, and input feature pages.