Google Professional Machine Learning Engineer Question 144
Single answerGoogle Cloud PlatformYou are a machine learning engineer tasked with building a custom model using Jupyter notebooks on Google Cloud. The model training requires access to GPUs for faster computation, and the dataset is stored in a BigQuery table. Which Jupyter backend should you choose to efficiently integrate with Google Cloud resources and utilize the required hardware acceleration?
- A
Vertex AI Workbench (Managed Notebooks) with a GPU-enabled instance
- B
Vertex AI Workbench (User-Managed Notebooks) with a CPU-only instance
- C
Google Colab with free-tier runtime
- D
A self-managed Jupyter notebook instance on a Compute Engine virtual machine (VM)
Show answer and explanation
Correct answer: A
Explanation
The most efficient and scalable choice is Vertex AI Workbench (Managed Notebooks) with a GPU-enabled instance. This option provides seamless integration with Google Cloud services such as BigQuery, supports GPU acceleration for faster computation, and simplifies resource management, making it ideal for machine learning tasks on Google Cloud.
- A. Correct.
Vertex AI Workbench (Managed Notebooks) with a GPU-enabled instance is the optimal choice because it provides tight integration with Google Cloud services like BigQuery, supports GPU acceleration, and simplifies resource management for machine learning workloads.
- B. Incorrect.
Vertex AI Workbench (User-Managed Notebooks) with a CPU-only instance does not meet the requirement for GPU acceleration, making it a suboptimal choice for the given scenario.
- C. Incorrect.
Google Colab with free-tier runtime does not guarantee GPU availability and lacks seamless integration with Google Cloud services like BigQuery for large-scale data processing.
- D. Incorrect.
A self-managed Jupyter notebook instance on a Compute Engine VM requires significant manual setup for GPU enablement and integration with Google Cloud services, making it less efficient compared to Vertex AI Workbench.