Sagemaker-python-sdk: How to create a hyperparameter tuning step in SageMaker pipeline?

[dennis-qian]

Describe the feature you'd like
I am trying to use the latest SageMaker Python SDK (v2.23.0) to implement a SageMaker pipeline that includes a hyperparameter tuning job. However I didn't see anything in module sagemaker.workflow.steps or sagemaker.workflow.step_collections that I can use. There is a TrainingStep class but it's not for HPO.

How would this feature be used? Please describe.
The new Step class will be used to do hyperparameter tuning in a SageMaker pipeline.

Describe alternatives you've considered
N/A

Additional context
N/A

[dennis-qian]

Is there any plan of including this feature request into a future release?

[supertzark]

how to get this implemented on pipeline. at least make some document to do this .

[wangweixun]

Hi there, thanks for your interests in SageMaker Pipelines.

We do not have a specific plan to release this feature right now. However your feedback is very important to us. We are always reprioritize our roadmap based on customer input.

[melzareix]

Hi, This feature is really important and necessary for us and it seems very odd not to be included in SM Pipelines where it's in the Step Functions Workflow API

[supertzark]

please check my medium book I have mentioned a method to implement hyperparameteron sm pipeline.... AWS please get us the feature as well..

https://codeloki.medium.com/understanding-aws-sagemaker-studio-with-model-building-training-and-deployment-along-with-the-12287ac56188