Amazon SageMaker customers can now select G4 and R5 instances when deploying their machine learning models for real-time inferences. Amazon SageMaker is a modular, fully-managed platform that enables developers and data scientists to quickly and easily build, train, and deploy machine learning models at any scale. G4 instances are optimized for machine learning application deployments such as image classification, object detection, recommendation engines, automated speech recognition and language translation that need access to low level GPU software libraries. R5 instances are memory optimized instances with EBS-optimized burst support, enabling customers to reduce costs by right-sizing instances based on their memory and compute requirements.
from Recent Announcements https://aws.amazon.com/about-aws/whats-new/2019/10/amazon-sagemaker-supports-g4-r5-instances-real-time-inference/