In this section, we will turn our attention from real-time predictions to batch predictions. To address the need to deploy trained models in offline mode, SageMaker offers Batch Transform. Batch Transform is a newly released high-performance and throughput feature where inferences can be obtained for the entire dataset. Both the input and output data is stored in an S3 bucket. The Batch Transform service manages the necessary compute resources to score the input data, given the trained model.
This following diagram shows how the Batch Transform service works:
In the preceding diagram, ...
