Professional Machine Learning Engineer

By Google · Question Mode

QuestionQ55

Automating and orchestrating ML pipelines

You are creating a custom image-classification model and intend to use Vertex AI Pipelines to implement end-to-end training. Your dataset contains images that must be preprocessed before they can train the model. The preprocessing includes:

  • resizing the images,
  • converting them to grayscale, and
  • extracting features.

You have already implemented Python functions for the preprocessing tasks. Which components should you use in the pipeline?

Explanation

dsl.component packages the existing Python preprocessing functions as pipeline components, and dsl.ParallelFor can run those preprocessing operations across the image collection. CustomTrainingJobOp runs the custom model-training workload in Vertex AI. Google Cloud documents CustomTrainingJobOp as the component for custom training and demonstrates creating Python pipeline components with @dsl.component.

Learn more

Community Discussion

No comments yet. Be the first to start the discussion!