AWS Certified Machine Learning – Specialty — Question 299

A machine learning (ML) specialist is using the Amazon SageMaker DeepAR forecasting algorithm to train a model on CPU-based Amazon EC2 On-Demand instances. The model currently takes multiple hours to train. The ML specialist wants to decrease the training time of the model.

Which approaches will meet this requirement? (Choose two.)

Answer options

• A. Replace On-Demand Instances with Spot Instances.
• B. Configure model auto scaling dynamically to adjust the number of instances automatically.
• C. Replace CPU-based EC2 instances with GPU-based EC2 instances.
• D. Use multiple training instances.
• E. Use a pre-trained version of the model. Run incremental training.

Correct answer: C, D

Explanation

Amazon SageMaker DeepAR utilizes recurrent neural networks (RNNs), which train significantly faster when accelerated by GPU-based EC2 instances compared to CPU-based instances. Additionally, DeepAR supports distributed training, so utilizing multiple training instances will parallelize the process and further decrease training time. Spot instances only optimize costs, while auto-scaling is intended for hosting endpoints rather than training.