A machine learning engineer wants to build a reusable workflow that includes data preparation, training, validation, and registration as separate steps with defined inputs and outputs. What should be implemented?

Question

Accepted Answer

A. A machine learning pipeline built from reusable components. Azure Machine Learning pipelines orchestrate multi-step workflows, and components provide reusable versioned building blocks. Option d is wrong because an idle shutdown schedule controls cost on a development machine and does not orchestrate a training workflow.

A machine learning engineer wants to build a reusable workflow that includes data preparation, training, validation, and registration as separate steps with defined inputs and outputs. What should be implemented?

Related Questions

Discussion