nerdexam
Microsoft

DP-100 · Question #302

You train and register a machine learning model. You create a batch inference pipeline that uses the model to generate predictions from multiple data files. You must publish the batch inference…

The correct answer is B. Azure Machine Learning compute cluster. The correct answer is B: Azure Machine Learning compute cluster. A batch inference pipeline processes large volumes of data files asynchronously on a schedule, which requires scalable compute that can spin up multiple nodes in parallel and scale down to zero when idle. An Azure…

Train and deploy models

Question

You train and register a machine learning model. You create a batch inference pipeline that uses the model to generate predictions from multiple data files. You must publish the batch inference pipeline as a service that can be scheduled to run every night. You need to select an appropriate compute target for the inference service. Which compute target should you use?

Options

  • AAzure Machine Learning compute instance
  • BAzure Machine Learning compute cluster
  • CAzure Kubernetes Service (AKS)-based inference cluster
  • DAzure Container Instance (ACI) compute target

How the community answered

(20 responses)
  • A
    5% (1)
  • B
    75% (15)
  • C
    15% (3)
  • D
    5% (1)

Explanation

The correct answer is B: Azure Machine Learning compute cluster. A batch inference pipeline processes large volumes of data files asynchronously on a schedule, which requires scalable compute that can spin up multiple nodes in parallel and scale down to zero when idle. An Azure ML compute cluster is designed exactly for this workload-it autoscales based on demand and is cost-efficient for scheduled batch jobs. A compute instance (A) is a single-node VM suited for development. An AKS inference cluster (C) is optimized for low-latency real-time (online) inference, not batch. ACI (D) is used for lightweight, short-lived real-time inference, not large-scale batch processing.

Topics

#Batch inference#Compute targets#Azure Machine Learning#Model deployment

Community Discussion

No community discussion yet for this question.

Sign up to join the discussion
Full DP-100 Practice