Artifact Schemas

This module provides a framework for defining reusable dumpers/loaders for custom model/dataset classes.

Schemas can be used standalone or registered in the tags of MLflow entities such as experiments, runs, models and model versions. In the second case, artifacts associated to those entities can be saved/loaded/verified just by providing the alias of a registered schema.

Contents:

• Framework
• Tags
• Helpers
• Artifact Specs
• Artifacts Catalog