This message was deleted.

Could you tell me more about what you had in mind? From reading elastic inference guide and document. I think Elastic inference is additional GPU resources that you can attach on EC2, sagemaker, and ECS(using AWS deep learning image). Right now, it is supports tensorflow, pytorch and mxnet by using AWS's own versions of those frameworks. You will need to install the enhanced framework package from AWS to make the inference with BentoML run on Elastic Inference. You can either use the AWS deep learning image as a base to build image from or download the required package from s3.

It's pricing is cheaper than alternatives, that's why we thought about using it