Answer-first summary for fast verification
Answer: AmlCompute
The correct answer is C (AmlCompute) because it is specifically designed for distributed processing and batch inference scenarios in Azure Machine Learning. AmlCompute provides managed compute infrastructure that can scale to handle parallel processing requirements of ParallelRunStep efficiently. The community discussion confirms this with 100% consensus on option C, and detailed explanations note that AmlCompute supports various VM sizes for scaling workloads. Other options are less suitable: A (BatchCompute) is not a valid Azure ML compute target class, B (AdlaCompute) is for Azure Data Lake Analytics, not ML workloads, and D (AksCompute) is optimized for real-time inference with low latency, not batch processing, making it unnecessarily costly and complex for this scenario.
Author: LeetQuiz Editorial Team
Ultimate access to all questions.
You create and register a model in an Azure Machine Learning workspace. You plan to implement a batch inference pipeline using the Azure Machine Learning SDK. This pipeline will use a ParallelRunStep to score input data with the registered model. You must specify a value for the compute_target setting of the ParallelRunConfig for the pipeline step.
Which class should you use to create the required compute target?
A
BatchCompute
B
AdlaCompute
C
AmlCompute
D
AksCompute
No comments yet.