Amazon SageMaker Batch Transform is a high throughput, asynchronous batch inference method for Amazon SageMaker. It’s ideal for scenarios where you need to run predictions on large datasets and don’t need sub-second latency1.

How Batch Transform Works

• Batch Transform partitions the Amazon S3 objects in the input by key and maps Amazon S3 objects to instances1.
• When a batch transform job starts, SageMaker initializes compute instances and distributes the inference or preprocessing workload between them1.
• It can preprocess datasets to remove noise or bias that interferes with training or inference from your dataset1.

Benefits of Batch Transform

• It can get inferences from large datasets1.
• It can run inference when you don’t need a persistent endpoint1.
• It can associate input records with inferences to assist the interpretation of results1.

Limitations of Batch Transform

• It doesn’t combine mini-batches from different input files to comply with the MaxPayloadInMB limit1.
• If you have one input file but initialize multiple compute instances, only one instance processes the input file and the rest of the instances are idle1.

Amazon SageMaker Processing

Amazon SageMaker Processing Jobs is a feature of Amazon SageMaker that allows you to easily run your preprocessing, postprocessing and model evaluation workloads on fully managed infrastructure.

How SageMaker Processing Works

• A Processing Job would need to be kicked off before the Batch Transform Job2.
• Once the Processing Job has completed you can then kick off the Batch Transform Job with the output of the Processing Job as input to the Batch Transform Job2.

Benefits of SageMaker Processing

• It allows you to easily run your preprocessing, postprocessing and model evaluation workloads on fully managed infrastructure.
• It can be used to process data for SageMaker Batch Inferencing2.