Applied Machine Learning for Healthcare and Life Sciences Using AWS by Ujjwal Ratan

Author:Ujjwal Ratan
Language: eng
Format: epub
Publisher: Packt
Published: 2022-11-15T00:00:00+00:00

Understanding Asynchronous Inference on SageMaker

A SageMaker asynchronous endpoint executes model inference by queuing up the requests and processing them asynchronously from the queue. This option is well suited for scenarios where you need near-real-time latency in your predictions and your payload size is really large (up to 1 GB), which causes long processing times. Unlike a real-time persistent endpoint, SageMaker Asynchronous Inference spins up infrastructure to support inference requests and scales it back to zero when there are no requests, thereby saving costs. It uses the simple notification service (SNS) to send notifications to the users about the status of the Asynchronous Inference. To use SageMaker Asynchronous Inference, you define the S3 path where the output of the model will be stored and provide the SNS topic that notifies the user of the success or error. Once you invoke the asynchronous endpoint, you receive a notification from SNS on your medium of subscription, such as email or SMS. You can then download the output file from the S3 location you provided while creating the Asynchronous Inference configuration.

Download

Copyright Disclaimer:
This site does not store any files on its server. We only index and link to content provided by other sites. Please contact the content providers to delete copyright contents if any and email us, we'll remove relevant links or contents immediately.