Replace SKLearn with supported code

Question

Hello,
Any ideas on how to rewrite this code to supported Python code?
This code in a demo does not run well:

WARNING:azureml.train.sklearn:'SKLearn' estimator is deprecated

from azureml.train.sklearn import SKLearn

estimator = SKLearn(source_directory='./Scripts',
compute_target=compute_target,
entry_script='train.py',
inputs=[tabular.as_named_input('training')],
pip_packages=['azureml-dataprep[fuse,pandas]','joblib==0.14.1','azureml-interpret','azureml-contrib-interpret','matplotlib','scikit-learn==0.22.1','seaborn'])

run = experiment.submit(estimator)
run

Answer 1

Hello,

Thanks for reaching out to us,"SKLearn" estimator classes has been deprecated in favor of using "ScriptRunConfig" to configure experiment runs.

Here I would recommend you to follow the document to see how to run your scikit-learn training scripts with Azure Machine Learning.
https://learn.microsoft.com/en-us/azure/machine-learning/how-to-train-scikit-learn?view=azure-ml-py

This is a good guidance for customer to migrate from SKLearn Estimator to ScriptRunConfig.

Hope this helps!

Regards,
Yutong

-Please kindly accept the answer if you feel helpful, thanks.