Saving data transformer for AzureML model

Question

I have an Azure Machine Learning training pipeline that uses scikit-learn's MinMaxScaler in the pre-processing step. What is the best way to save the MinMaxScaler (or any other transformer) so that it can later be used during inference. Ideally I would like to link the transformer with the model.

One idea that I had was to save the pickle file of the scaler in the Blob storage, then add the path to it as a property of the model. Is there an easier way of achieving this?

Answer 1

When we create an AutoML, after getting the results, we need to create the registration for the model. Click on the register model to get the properties of the model and about complete information of the model.

We are getting the properties here and use these properties to get the model information. By default we can get the model to be connected to the blobstorage.