Exam DP-100 topic 3 question 85 discussion

Create and run a training pipeline.. Create and run a real-time inference pipeline.. Deploy a real-time endpoint..

Why create another compute instance? (it clearly states that he already has one) I would remove that and move everything up and add the deploy the end point https://docs.microsoft.com/en-us/azure/machine-learning/tutorial-designer-automobile-price-deploy

D-C-B is correct for me. We can use pre-built compute cluster to run AML Designer, so no need to create a new compute instance.

The correct sequence of actions for creating a web service to consume the classification model would be: D-Create and run a training pipeline that prepares the data and trains a classification model on the compute cluster. First, you need to use the data in your tabular dataset to train a classification model. This is done on your compute cluster and involves creating a pipeline that prepares the data for machine learning and then trains the model. C-Create and run a real-time inference pipeline on the compute cluster. After your model is trained, you need to create an inference pipeline. This pipeline will take new data, run it through your model, and output the model's prediction. Since you want the client applications to get an immediate response, you should use a real-time inference pipeline. B-Deploy a real-time endpoint on the inference cluster. After your inference pipeline is created, you deploy it as a web service on the AKS inference cluster. This endpoint will allow client applications to send new data and receive predictions in real-time.

The correct sequence of actions to use the Azure Machine Learning designer to create a web service for a classification model would be: Create and run a training pipeline that prepares the data and trains a classification model on the computer cluster. Use the automated ML user interface to train a classification model on the compute cluster. Create and run a real-time inference pipeline on the compute cluster. Deploy a real-time endpoint on the inference cluster. The actions "Create and start a Compute Instance" and "Create and run a batch inference pipeline on the compute cluster" are not necessary for creating a web service to consume a classification model.

Similar question on exam 18-06-22

On 20Feb2022

Question says workspace that contains a CPU-based compute cluster, so we don't need to create a Compute Instance. Questions says need to use the Azure Machine Learning designer to create a web service. -> Use the Automated ML user interference pipeline -> create and run a real-time inference pipeline on the computer cluster -> deploy a real-time endpoint on the inference cluster

Given answer is wrong, you definitely need to deploy the model to consume it. We can run the pipeline on a compute cluster as well, I do not see a need to create and start a compute Instance again.

Create and run a training pipeline Create and run a real-time inference pipeline. Deploy a real-time endpoint.

we do not need a compute instance, we already have CPU-based compute cluster, I agree with the answer of @kty Create and run a training pipeline Create and run a real-time inference pipeline. Deploy a real-time endpoint.

On Exam 01 Oct 2021

Create and start a compute instance -> create and run a real-time inference pipeline on the computer cluster -> deploy a real-time endpoint on the inference cluster

On exam 2021/7/10