Databricks Model Registry Management Quiz

1. Which API or interface allows you to programmatically interact with the Model Registry?

Only the Databricks web UI

MLflow Python API and REST endpoints

Only SQL queries

Only command-line tools

2. What is the purpose of model stage transitions in the Registry workflow?

To randomly reorganize models

To control model lifecycle progression from development through production with quality gates

To increase model accuracy

To reduce computational costs

3. Can you add custom metadata to a model in the Model Registry?

No, only predefined fields are allowed

Yes, through tags and custom key-value properties

Only through the Python API, not the UI

No, metadata is automatically generated

4. What is the relationship between MLflow Tracking and Model Registry?

They are completely independent systems

Tracking logs experiments; Registry manages the lifecycle of logged models

Registry only works with models not tracked

They serve identical purposes

5. How does the Model Registry facilitate collaboration among data scientists?

By restricting access to models

By providing a centralized repository where models can be discovered, shared, and evaluated

By requiring models to be stored locally only

By preventing concurrent access to models

6. Which of the following is a best practice for model versioning in the Registry?

Deleting old versions to save storage space

Using descriptive tags and maintaining detailed change logs for each version

Keeping all versions in Production simultaneously

Avoiding version numbering

7. What is a registered model in the Databricks Model Registry?

A collection of model versions with the same name and associated metadata

A single training run result

A database table containing model parameters

A Python class definition

8. How can you implement CI/CD workflows with the Databricks Model Registry?

By manually uploading models daily

By automating model transitions between stages based on test results and approvals

By storing models in external databases only

By disabling version control

9. In the Model Registry, what does the 'None' stage indicate?

A model version that is active and ready for use

A model version with no stage assigned; typically a new or experimental version

A model that cannot be accessed

A model in Production stage

10. What is model lineage in the context of Databricks Model Registry?

The sequence of model versions and their dependencies on training data and code

The alphabetical order of model names

The geographic location where the model is deployed

The cost associated with model training

11. What is the primary purpose of the Databricks Model Registry?

To store raw training data in a centralized location

To manage the complete ML model lifecycle including versioning, staging, and deployment

To compile Python code into executable binaries

To monitor cluster performance metrics in real-time

12. What happens when you archive a model version?

The version is permanently deleted from the system

The version is marked as inactive but remains accessible for reference

The version is automatically moved to Production

The version is encrypted and cannot be viewed

13. How does the Model Registry support model governance and compliance?

By automatically deleting models after 30 days

By providing audit trails, change logs, and access controls

By encrypting model data only

By limiting model versions to five per model

14. What metadata can be attached to a model version in the Model Registry?

Only the model accuracy metric

Tags, descriptions, and custom key-value pairs

Only the training dataset name

Only the model file size

15. Can multiple model versions exist in the Production stage simultaneously?

Yes, unlimited versions can be in Production at once

No, only one version can be in Production at a time

Yes, but maximum of three versions

Only if they have different model types

16. What is the Staging stage typically used for in the model lifecycle?

Initial model training and experimentation

Testing model changes before promoting to Production

Archiving old models no longer in use

Storing model source code

17. Which MLflow component does Databricks leverage for Model Registry functionality?

MLflow Tracking

MLflow Projects

MLflow Model Registry

MLflow Serving

18. How do you transition a model version to the Production stage?

By automatically deploying it after training completes

By explicitly changing the stage through the Model Registry UI or API

By renaming the model artifact

By increasing the model's accuracy score above a threshold

19. In the Model Registry, what does a model version represent?

A snapshot of the model at a specific point in time with unique metadata

Only the hyperparameters used in training

The computational resources allocated to the model

The number of times the model has been deployed

20. Which of the following is NOT a standard model stage in the Databricks Model Registry?

None (archived)

Staging

Production

Testing

Databricks ML Associate Model Registry Management Quiz