Databricks Machine Learning Associate Question 646
Select 4You have trained a machine learning model in Databricks and registered it in the MLflow Model Registry. You need to deploy the model as a REST endpoint for real-time inference. Which of the following steps are required to accomplish this task?
- A
Use MLflow to transition the model to the 'Production' stage in the Model Registry
- B
Deploy the model to a Databricks serving endpoint
- C
Configure the model's serving endpoint to accept requests and return predictions
- D
Write a batch inference job to test the model and confirm its performance
- E
Use the REST API URL from the Databricks serving endpoint to send real-time queries
Show answer and explanation
Correct answers: A, B, C, E
Explanation
To deploy a model for real-time inference in Databricks, you must transition the model to 'Production' in the MLflow Model Registry, deploy it to a serving endpoint, configure the endpoint for accepting requests, and use the provided REST API URL for querying. Batch inference is unrelated to real-time deployment.
- A. Correct.
Correct. Transitioning the model to the 'Production' stage ensures it is ready for deployment and indicates its stability for serving.
- B. Correct.
Correct. Deploying the model to a Databricks serving endpoint is essential to make it available for real-time inference.
- C. Correct.
Correct. Configuring the serving endpoint to handle requests and return predictions is a necessary step for real-time inference.
- D. Incorrect.
Incorrect. A batch inference job is not required for deploying the model as a REST endpoint for real-time inference. Batch inference serves a different purpose.
- E. Correct.
Correct. The REST API URL is used to send real-time queries to the deployed model.