1

I just starting learning Machine learning with azure. I have read a few docs about the difference between scriptrunconfig, runconfig and estimator. I didn't understand much about them in any of the doc. please help me understand them. when do we use each one and why???

sheshank k
  • 11
  • 2

1 Answers1

1

Although all three mentioned methods are used for configuring a training job via the Azure ML SDK, please know that ScriptRunConfig is the recommended option going forward to simplify the job configuration process.

As clarified in the docs:

Estimators are deprecated with the 1.19.0 release of the Python SDK. You should also generally avoid explicitly instantiating a RunConfiguration object yourself, and instead configure your job using the ScriptRunConfig class.

Also, Azure Machine Learning documentation and samples have been updated to use ScriptRunConfig for job configuration and submission to address this ambiguity and inconsistency with multiple options.

Here is a helper doc to guide through the migration process: Migrating from Estimators to ScriptRunConfig

Bhargavi Annadevara
  • 4,923
  • 2
  • 13
  • 30
  • Unfortunately, Microsoft hasn't updated all the trainings documents. For example, you can see Microsoft still uses Runconfig in the mslearn-dp100 repository (08 - Create a Pipeline) as follows: pipeline_run_config = RunConfiguration() – user3665906 Apr 18 '22 at 00:08
  • @user3665906 Please report it as a [new issue](https://github.com/MicrosoftLearning/mslearn-dp100/issues/new) for the repo owners to action on. :) – Bhargavi Annadevara Apr 19 '22 at 04:52