AWS Certified Machine Learning Engineer Associate · Free Practice Question Medium
Question 7
A company is training a deep learning model with billions of parameters, which is too large to fit on a single machine. They decide to use model parallelism on Amazon SageMaker.
What key feature of SageMaker makes model parallelism easier to implement, and how does it work?
-
A
SageMaker’s Model Parallelism Library automatically splits the model across multiple GPUs or machines, based on the model architecture and available memory.
-
B
SageMaker’s built-in algorithms dynamically adjust the batch size to ensure the model can be trained on multiple machines.
-
C
SageMaker handles model parallelism by dividing the dataset across multiple machines and averaging the model gradients.
-
D
SageMaker Model Monitor automatically manages and optimizes the model splitting process across multiple GPUs.
Reveal correct answer
Correct answer: A
A.
SageMaker’s Model Parallelism Library simplifies the process by automatically splitting the model across multiple GPUs or machines based on the architecture and memory constraints, making it easy to handle large models without manual intervention.
B.
Adjusting the batch size is not related to model parallelism, which involves splitting the model itself, not just tuning batch size.
C.
This describes data parallelism, where the dataset is divided, not the model.
D.
SageMaker Model Monitor is used for monitoring model performance and data drift, not for managing model parallelism.
Discussion
Think the marked answer is wrong, or have a better explanation? Share it below — comments appear after review.
