Databricks Generative AI Engineer Associate Question 212
Select 3You are working on a Databricks project where you’ve trained a machine learning model for customer churn prediction, and you want to register this model in Unity Catalog using MLflow. Which of the following steps are required to successfully register the model in Unity Catalog?
- A
Set up a Unity Catalog-enabled Databricks workspace with access permissions configured.
- B
Use the
mlflow.register_model()function specifying the Unity Catalog model registry URI. - C
Ensure the model is saved in the MLflow format before registering it in Unity Catalog.
- D
Enable Delta Sharing on the Unity Catalog instance to facilitate model registration.
- E
Specify a unique model name in Unity Catalog during registration.
Show answer and explanation
Correct answers: A, C, E
Explanation
To register a model in Unity Catalog using MLflow, you need to ensure the workspace has Unity Catalog enabled with proper permissions, the model is saved in the MLflow format, and a unique model name is provided. These steps are essential for successfully integrating models into Unity Catalog's centralized governance system.
- A. Correct.
Correct: A Unity Catalog-enabled Databricks workspace is required as Unity Catalog centralizes governance for models. You also need proper permissions to register models.
- B. Incorrect.
Incorrect: MLflow does not have a function
mlflow.register_model(); instead, MLflow's Model Registry API is used for operations like registering models. - C. Correct.
Correct: Models must be saved in the MLflow format to ensure compatibility with MLflow APIs and Unity Catalog registration.
- D. Incorrect.
Incorrect: Delta Sharing is not required for registering models in Unity Catalog. Delta Sharing is used for sharing data securely, not for model registration.
- E. Correct.
Correct: Unity Catalog requires a unique model name to avoid conflicts in the centralized registry.