a13zen

Anny chance this could be pushed through? Seems like the linked PR for this is almost ready?

Any new updates on this?

@shivlaks, one common ask is for ml pipelines to have all input/outputs for each step stored in the same location on s3 for lineage reasons. In order to do that,...

We also have an utility function for finding the trained model when using such dynamic paths ```python def trained_model_path(*, suffix): return f"States.Format('{bucket_uri}/{{}}/{suffix}/{{}}/output/', $$.Execution.Name, $.training_job_name)" ``` Since training job name is...