Reduce Inferencing Cost by Up to 90% Using Amazon Elastic Inference and Amazon EC2 Spot Instances

Level: 300

In most deep learning applications, making predictions using a trained model with a process known as inference, can drive as much as 90% of the compute costs of the application. While Amazon Elastic Inference solves this problem by allowing you to attach just the right amount of GPU-powered inference acceleration to any EC2 and reduce costs on Inferencing, using EC2 Spot with Elastic Inference can further reduce your compute costs up to 90%. In this tech talk, we'll discuss cost optimization of Amazon Elastic Inference running with Amazon EC2 Spot instances and walk through the best practices by using Cloudformation and launch templates for build automation.