Databricks Machine Learning Associate Question 489
Single answerYou are managing machine learning models in Databricks Model Registry. A new model version (Version 5) has been trained and registered as a challenger model to replace the current champion model (Version 3). You want to promote the challenger model to become the new champion using aliases. What steps should you take to perform this action?
- A
Assign the alias 'champion' to Version 5 and remove the 'champion' alias from Version 3.
- B
Delete Version 3 from the Model Registry and assign the 'champion' alias to Version 5.
- C
Assign the alias 'challenger' to Version 5 and keep Version 3 with the 'champion' alias.
- D
Create a new alias named 'new_champion' for Version 5 and leave Version 3 unchanged.
Show answer and explanation
Correct answer: A
Explanation
In Databricks Model Registry, aliases such as 'champion' and 'challenger' are used to label specific versions of a model. To promote a challenger model to champion, you must reassign the 'champion' alias to the new version (Version 5) and remove it from the previous champion model (Version 3). This ensures the alias accurately reflects the current champion model.
- A. Correct.
Correct: To promote a challenger model (Version 5) to a champion, you must assign the 'champion' alias to it and remove the alias from the current champion (Version 3). This ensures the correct alias points to the desired model version.
- B. Incorrect.
Incorrect: Deleting the current champion model (Version 3) is unnecessary. Aliases can be updated without deleting models.
- C. Incorrect.
Incorrect: Keeping Version 3 as the 'champion' while assigning 'challenger' to Version 5 does not promote the challenger to champion. The 'champion' alias must point to Version 5.
- D. Incorrect.
Incorrect: Creating a new alias like 'new_champion' does not replace the 'champion' alias. The alias 'champion' must be reassigned to Version 5 to complete the promotion.