Question 34
Domain 4: Deploy and operationalize machine learning solutionsYou have used the Python SDK for Azure Machine Learning to create a pipeline that trains a model. What do you need to do so that a client application can invoke the pipeline through an HTTP REST endpoint?
Correct answer: B
Explanation
Azure Machine Learning pipelines must be published to expose them as a REST endpoint. Publishing creates an HTTP endpoint that a client application can call to run the pipeline remotely.
Why each option is right or wrong
A. Create an inference cluster compute target.
B. Publish the pipeline.
Azure Machine Learning exposes a pipeline for remote invocation only after it is published as a pipeline endpoint; in the Python SDK this is done with the pipeline’s publish method, which creates a REST URI and an associated endpoint name. Until publication, the pipeline exists only as a workspace asset and cannot be called directly over HTTP by a client application.
C. Rename the pipeline to pipeline_name-production.