AWS Certified Machine Learning Specialty · Free Practice Question Medium
Question 11
A data science team is planning to train a series of models using XGBoost version 1.2 or newer on Amazon SageMaker for a predictive analytics project. Given the computational requirements of these models, which instance type would be the most appropriate for optimizing training performance and cost-efficiency on SageMaker?
-
A
Choose a P3 instance type
-
B
Use an R5 instance type.
-
C
Opt for a T3 instance type.
-
D
Select an M5 instance type.
Reveal correct answer
Correct answer: A
Explanation
Focus on the introduction of GPU support in XGBoost 1.2 and the resulting implications for instance type selection.
Correct choice: Choose a P3 instance type.
P3 instances offer GPU capabilities that significantly accelerate training times when using XGBoost 1.2 or newer with the gpu_hist option, balancing cost and efficiency better than CPU-only instances for this specific use case.
Incorrect choice: Select an M5 instance type.
While M5 instances are well-suited for CPU-based computations, they lack the GPU acceleration that XGBoost 1.2 can exploit through the gpu_hist parameter, making them less efficient for this scenario.
Incorrect choice: Opt for a T3 instance type.
T3 instances are designed for cost-efficiency in low to moderate workloads with occasional performance bursts but do not offer the GPU support needed for the optimized use of XGBoost 1.2's gpu_hist feature, leading to longer training times.
Incorrect choice: Use an R5 instance type.
R5 instances are optimized for memory-intensive applications and do not provide the GPU acceleration benefits required by XGBoost 1.2 when utilizing the gpu_hist parameter, making them a suboptimal choice for GPU-enabled machine learning tasks.
Discussion
Think the marked answer is wrong, or have a better explanation? Share it below — comments appear after review.
