Amazon SageMaker Batch Transform Assists TFRecord Format

Amazon SageMaker offers developer and data scientist the potential to build, train, and deploy machine learning models rapidly. Amazon SageMaker is a completely organized service which includes the whole machine learning workflow to label and assemble your data, select an algorithm, instruct the algorithm, tune and optimize it for deployment, make a forecast, and take action. A major feature within SageMaker is Batch Transform that enables you to run predictions on batch data. Now this Amazon SageMaker Batch Transform helps TFRecord format as a supported SplitType, allowing datasets to be divided by TFRecord boundaries. This appends to the list of supported formats covering RecordIO, CSV, and Text. TFRecord is a standard TensorFlow data format. It is a record-oriented binary file format, allowing well planeed storage and processing of sizeable datasets. Now TFRecord support is obtainable in every AWS regions where Amazon SageMaker is accessible. To get detail information, refer documentation and sample example