Amazon SageMaker Supports High Throughput Batch Transform Jobs for Non-Real Time Inferencing

Posted on: Jul 17, 2018

Amazon SageMaker now supports fully-managed high-throughput batch transform jobs for non-real time inferencing. Existing machine learning models developed on Amazon SageMaker can work seamlessly with this new capability without any changes.

Till now, processing batches of data for non-real time inferencing needed to be done by resizing large datasets into smaller chunks of data and managing real-time endpoints. With the new Batch Transform capability, customers can process batch jobs within Amazon SageMaker, through a simple API call irrespective of the data set sizes. Batch Transform jobs can be done on a range of data sets, from petabytes of data to very small data sets. SageMaker manages the provisioning of resources at the start of the job and releases them when the job is complete. The output of the batch transform jobs is stored in the S3 bucket chosen by the user.

Amazon SageMaker Batch Transform is now available in the US East (N. Virginia), US East (Ohio), US West (Oregon), Europe (Ireland), Europe (Frankfurt), Asia Pacific (Tokyo), Asia Pacific (Seoul), and Asia Pacific (Sydney) AWS Regions. More information on Amazon SageMaker can be found here.