Databricks Machine Learning Professional Question 91
Select 3You are managing the lifecycle of a machine learning model in Databricks. Your team has decided to automate the process of transitioning models to the 'Archived' stage when they are no longer in use and delete models that have been archived for over 30 days. Which of the following steps would you include in your automation pipeline to achieve this?
- A
Use the Databricks REST API to list all model versions in the 'Archived' stage and filter those older than 30 days.
- B
Manually set a reminder to archive models after they are no longer in use.
- C
Implement a scheduled job in Databricks to transition unused models to the 'Archived' stage.
- D
Use the Databricks REST API to delete model versions that have been in the 'Archived' stage for more than 30 days.
- E
Manually delete all models in the 'Archived' stage using the Databricks web interface.
Show answer and explanation
Correct answers: A, C, D
Explanation
To automate the model lifecycle, you need to identify models in the 'Archived' stage that meet the deletion criteria, transition unused models to the 'Archived' stage programmatically, and remove old models programmatically using the Databricks REST API. Manual interventions do not align with the goal of automation.
- A. Correct.
This step is essential to identify which model versions in the 'Archived' stage are older than 30 days, enabling their deletion.
- B. Incorrect.
This is not an automated approach and would go against the goal of lifecycle automation.
- C. Correct.
A scheduled job in Databricks can programmatically transition models to the 'Archived' stage when they are no longer in use, aligning with automation goals.
- D. Correct.
This step ensures old archived models are automatically removed, completing the lifecycle automation process.
- E. Incorrect.
Manually deleting models through the web interface is not automated and would not fulfill the requirements for automation.