You use Azure Machine Learning designer to publish an inference pipeline. Which two parameters should you use to consume the pipeline? Each correct…

When an Azure Machine Learning designer pipeline is published as a real-time inference endpoint, it is exposed as a web service. To consume this service, a client application must know its location and have the necessary credentials to access it. The REST endpoint provides the unique URL (Uniform Resource Locator) for the service. The authentication key (either primary or secondary) serves as a password, which must be included in the request header to authenticate the client and authorize the call. Together, these two parameters allow a client to securely send data to the pipeline and receive a prediction.

A. the model name: The model name is an internal identifier for the registered model asset within the Azure Machine Learning workspace, not the public address used to call the service.

B. the training endpoint: The published pipeline is an inference endpoint used for making predictions, not a training endpoint. Training and inference are distinct processes with separate pipelines.

1. Microsoft Learn

Azure Machine Learning Documentation. "Consume an endpoint from Azure Machine Learning studio." In the "Consume the endpoint" section

it explicitly states: "To consume the endpoint

you'll need the REST URI and the primary key." The page provides code examples in Python

C#

and R

all of which utilize variables for the endpoint URL and the authentication key.

Source: https://learn.microsoft.com/en-us/azure/machine-learning/how-to-consume-endpoint?tabs=python%2Creal-time (Section: "Consume the endpoint")

2. Microsoft Learn

Azure Machine Learning Documentation. "Tutorial: Deploy a machine learning model with the designer." In Step 3

"Test the real-time endpoint

" the instructions guide the user to the "Consume" tab of the deployed endpoint. It states: "You can find the REST endpoint's URI and the primary key to the endpoint here. You need these two pieces of information to consume the endpoint."

Source: https://learn.microsoft.com/en-us/azure/machine-learning/tutorial-designer-automobile-price-deploy?tabs=deploy#test-the-real-time-endpoint (Paragraph 1)