A machine learning team is training a large image classification model on Amazon SageMaker using Pipe input mode. Which data format is most efficient for streaming large volumes of image data in Pipe mode?

Question

Accepted Answer

C. RecordIO. RecordIO is a binary format optimized for SageMaker Pipe mode streaming. It packs records into binary sequences that SageMaker reads directly from S3 without downloading the entire dataset first, significantly improving training throughput for large datasets.

A machine learning team is training a large image classification model on Amazon SageMaker using Pipe input mode. Which data format is most efficient for streaming large volumes of image data in Pipe mode?

Related Questions

Discussion