Google Professional Machine Learning Engineer Question 141
Select 3Google Cloud PlatformYou are tasked with rapidly prototyping a machine learning model using Jupyter notebooks on Google Cloud. Which considerations should you prioritize to ensure an efficient and scalable prototyping process?
- A
Use pre-configured AI Platform Notebooks to avoid setting up dependencies manually.
- B
Store all intermediate outputs and datasets directly on the virtual machine's local disk for faster access.
- C
Leverage BigQuery to query large datasets directly into your notebook for preprocessing.
- D
Focus on using hardware accelerators such as GPUs or TPUs even if your prototype does not require heavy computation.
- E
Version your code and notebooks using a Git-based repository integrated with your development process.
Show answer and explanation
Correct answers: A, C, E
Explanation
Prototyping machine learning models in Jupyter notebooks requires leveraging Google Cloud's managed services like AI Platform Notebooks for dependency management, BigQuery for scalable data querying, and proper version control for reproducibility. Avoid practices that lead to potential data loss or unnecessary resource costs, such as storing data on local disks or using accelerators without justification.
- A. Correct.
Using pre-configured AI Platform Notebooks simplifies dependency management and ensures compatibility with Google Cloud services, which is crucial for rapid prototyping.
- B. Incorrect.
Storing data on the virtual machine's local disk is not recommended as it can be lost if the instance is stopped or terminated. Using persistent storage like Cloud Storage is a better practice.
- C. Correct.
BigQuery enables efficient querying of large datasets, eliminating the need for downloading massive files locally. This is highly beneficial for large-scale prototyping.
- D. Incorrect.
While hardware accelerators like GPUs or TPUs are useful for training deep learning models, they are unnecessary for lightweight or non-computationally intensive prototypes, leading to unnecessary costs.
- E. Correct.
Versioning your code and notebooks ensures traceability, reproducibility, and collaboration, which are vital for an organized prototyping workflow.