New Release of the Model Catalog in Data Science

This  new release of the model catalog is now available. It includes these enhancements: 

  • New model taxonomy metadata that lets data scientists document the use case, framework, and hyperparameters of their models.
  • Improvement to the model provenance metadata, including a reference to the model training resource (notebook sessions).
  • Support for custom metadata which lets data scientists document the context around their models including references to the conda environment used to train the model, the training and validation datasets, and so on. 
  • Support for documenting the model input feature vector and prediction schemas.
  • Model introspection tests that are run on the model artifact before the model is saved to the model catalog. Model introspection validates the artifact against a series of common issues and errors found with artifacts. These introspection tests are part of the model artifact code template that is included. 

Along with model catalog, a series of new conda environments with an updated version of Accelerated Data Science (ADS) are included in this release. ADS now includes the new model catalog features. ADS now automatically extracts model taxonomy, provenance, input and output schemas, and runs introspection tests automatically. Take a look at the ADS release notes for all of the changes.

We have created a few notebook examples showcasing the new model catalog features on our public Github repository. Take a look at our Data Science blog that introduces these new model catalog features.