search

aws / aws-step-functions-data-science-sdk-python

Step Functions Data Science SDK for building machine learning (ML) workflows and pipelines on AWS
Apache License 2.0
285 stars 87 forks source link

fix: Support placeholders for hyperparameters passed to TrainingStep #159

Closed ca-nguyen closed 3 years ago

ca-nguyen commented 3 years ago

Issue #, if available: #152

Description of changes: A breaking change was introduced in V2.2.0 where it was no longer possible to use Placeholders as input for hyperparameters in TrainingStep.

With this solution, TrainingStep hyperparameters will be compatible with Placeholders with the same behaviour as in V2.1.0: if a Placeholder is provided as input to hyperparameters in TrainingStep constructor, the estimator hyperparameters are overwritten

A feature request to be able to merge the Placeholder hyperparameters was made in #152, which will be addressed in another PR.

By submitting this pull request, I confirm that you can use, modify, copy, and redistribute this contribution, under the terms of your choice.

ca-nguyen commented 3 years ago

🎉 I think this issue is a fix and not a feat. had a couple other small comments.

Done!

Looks good, just some minor comments for clarity

Comments addressed :)

StepFunctions-Bot commented 3 years ago

AWS CodeBuild CI Report

Powered by github-codebuild-logs, available on the AWS Serverless Application Repository