Google Professional Machine Learning Engineer exam dumps

Google Professional Machine Learning Engineer practice question 386 of 522

Professional Machine Learning Engineer. Professional level, Google Cloud. Free question with the correct answer and a full explanation.

Google Professional Machine Learning Engineer Question 386

Select 3Google Cloud Platform

You are deploying a machine learning model as a REST API using Google Cloud Run. The model requires GPU acceleration for inference and needs to access private datasets stored in a Google Cloud Storage bucket. Which of the following steps are necessary to meet these requirements?

  1. A

    Ensure that the Cloud Run service is deployed in a region where GPUs are supported.

  2. B

    Enable GPU support in Cloud Run by selecting the appropriate machine type with GPU during deployment.

  3. C

    Grant the Cloud Run service account the 'Storage Object Viewer' role for the Google Cloud Storage bucket.

  4. D

    Use a custom Docker container to package the machine learning model and ensure the container runtime supports GPU drivers.

  5. E

    Configure VPC connector to allow the Cloud Run service to access private resources.

Show answer and explanation

Correct answers: C, D, E

Explanation

Cloud Run does not natively support GPUs, so GPU-based inference cannot be directly achieved. Instead, you can package the ML model in a custom Docker container with GPU support and deploy it to a service like Compute Engine or AI Platform if GPU is essential. However, for the scenario described, using a custom container, granting appropriate permissions to access private datasets, and configuring a VPC connector to access private resources are necessary steps to meet the requirements.

  • A. Incorrect.

    Cloud Run does not currently support GPUs as of October 2023, so selecting a GPU-supported region for deployment is not applicable.

  • B. Incorrect.

    Cloud Run does not allow the use of GPUs directly. GPU support is available in other services like Compute Engine and AI Platform, not in Cloud Run.

  • C. Correct.

    Granting the 'Storage Object Viewer' role to the Cloud Run service account ensures the service can access the private datasets stored in the Google Cloud Storage bucket.

  • D. Correct.

    Since Cloud Run does not natively support GPUs, a custom Docker container must be used to package the ML model, and it should include runtime dependencies, such as GPU drivers, if needed.

  • E. Correct.

    A VPC connector is required for Cloud Run to access private resources, such as a private Google Cloud Storage bucket.

Timed practice exam

Take a Google Professional Machine Learning Engineer practice test under exam conditions

60 questions in 120 minutes, drawn from this bank, with a score report and a per-question review when you finish.

Start timed exam

More Google Professional Machine Learning Engineer practice questions

  1. 387.Your team has developed a REST API-based machine learning model inference service. The service...
  2. 388.You are deploying a machine learning inference service that processes real-time predictions using a...
  3. 389.You are designing a machine learning workflow for a model that will process large-scale time-series...
  4. 390.You are designing a machine learning workflow to train and deploy a model using Google Cloud. Your...
  5. 391.You are tasked with building a machine learning workflow to preprocess data, train a model, and...
Back to the Google Professional Machine Learning Engineer exam dumps