aws / sagemaker-python-sdk

A library for training and deploying machine learning models on Amazon SageMaker
https://sagemaker.readthedocs.io/
Apache License 2.0
2.1k stars 1.14k forks source link

Transformer.transform() doesn't return job name #2858

Open shreeshaprabhu opened 2 years ago

shreeshaprabhu commented 2 years ago

Describe the feature you'd like

Get name of batch transform job after sagemaker.transformer.Transformer.transform() call, because the method does not return anything.

How would this feature be used? Please describe.

Job name will be used for bookkeeping purposes. Job name will be used to locate logs if debugging is required.

Describe alternatives you've considered

When transform method is executed, it sets the object field latest_transform_job. We are planning to use transformer. latest_transform_job.name to get the job name.

But the Transformer documentation doesn't mention anything about the field latest_transform_job. We are not sure if the field is safe for use or if the field will get removed/replaced in the future versions.

Additional context

navinsoni commented 1 year ago

@shreeshaprabhu This function returns pipeline step arguments in case the Transformer instance is built with PipelineSession, returning job_name with break existing customers.

martinRenou commented 1 year ago

@navinsoni it seems that in the current transform implementation, it never returns anything. I suspect returning the name of the transform job would not break any customer code?