Databricks Machine Learning Associate Question 635
Select 3You are tasked with deploying a custom machine learning model as a REST API endpoint using Databricks Model Serving. The model has been registered in the Databricks Model Registry. Which steps are required to successfully deploy the model to a model endpoint?
- A
Enable model serving for the Databricks workspace.
- B
Select the model version from the Databricks Model Registry and enable serving for it.
- C
Write a custom Flask application to serve the model.
- D
Test the deployed endpoint using the provided scoring URL.
- E
Configure the endpoint to use the Databricks runtime for inference.
Show answer and explanation
Correct answers: A, B, D
Explanation
To deploy a custom model to a model endpoint in Databricks, you need to ensure that model serving is enabled at the workspace level, select the desired model version from the Model Registry, and enable serving for it. Once deployed, testing the endpoint using the scoring URL is essential to validate its functionality. Additional custom application development or manual runtime configuration is not required, as Databricks manages these aspects automatically.
- A. Correct.
Enabling model serving for the Databricks workspace is a prerequisite to deploy models as endpoints. Without this, serving cannot be configured.
- B. Correct.
The model version must be selected from the Databricks Model Registry, and serving needs to be enabled for that specific version to deploy it as an endpoint.
- C. Incorrect.
Writing a custom Flask application is not required, as Databricks Model Serving automatically handles the deployment and serving process.
- D. Correct.
Testing the deployed endpoint using the provided scoring URL is an essential step to ensure the model endpoint is functioning as expected.
- E. Incorrect.
Configuring the endpoint to use the Databricks runtime for inference is not a manual step. Databricks automatically manages the runtime environment for serving models.