0

I have a customized score.py file which was generated within databricks but I didn't find a way to deploy it on a container.

I am using the mlflow.azureml, on my image creation I couldn't find how to specify the score.py in particular.

import mlflow.azureml

model_image, azure_model = mlflow.azureml.build_image(model_uri=model_uri, 
                                                      workspace=workspace,
                                                      model_name="my_model",
                                                      image_name="image_name",
                                                      description="Predicts",
                                                      synchronous=False)

Is there a way to specify the score.py using the lib?

Emerson R Orci
  • 85
  • 1
  • 6

1 Answers1

0

Specifying custom score file is not currently supported through the mlflow.azureml APIs, it auto-generates it behind the scenes.

You can switch to using the Azure ML Python SDK Model class which has the deploy method with your own score.py as argument.

Roope Astala - MSFT
  • 756
  • 3
  • 5