Databricks Machine Learning Associate Question 53
Single answerYou have trained multiple models using Databricks' AutoML functionality to predict housing prices. Upon completion, AutoML has selected the best-performing model based on the evaluation metric. How can you locate the source code for this best model in the Databricks workspace?
- A
Navigate to the AutoML Experiment page and click on the 'Best Trial' link to access the source code notebook.
- B
Download the model as a Python file from the AutoML Experiment page.
- C
Check the Databricks Repos section for an automatically generated notebook containing the best model's source code.
- D
Open the MLflow tracking page for the AutoML experiment, locate the best model trial, and access the linked notebook for the trial.
Show answer and explanation
Correct answer: A
Explanation
Databricks AutoML generates a notebook for each trial, including the best-performing model. To locate the source code for the best model, you need to navigate to the AutoML Experiment page and click on the 'Best Trial' link, which redirects you to the notebook containing the implementation for the best model. This is the most direct and accurate way to locate the source code.
- A. Correct.
Correct. The AutoML Experiment page in Databricks provides a direct link to the source code notebook for the best trial under the 'Best Trial' section.
- B. Incorrect.
Incorrect. Databricks AutoML does not provide an option to directly download the model source code as a Python file.
- C. Incorrect.
Incorrect. AutoML-generated notebooks are not automatically stored or synced in the Databricks Repos section.
- D. Incorrect.
Incorrect. While the MLflow tracking page is useful for viewing experiment details, the source code for the best trial is not linked directly from there. Instead, it is accessible from the AutoML Experiment page.