none
How to test the deployed model: autoML. RRS feed

  • Question

  • Hi,

    I created a model using Azure automated machine learning (classification tutorial) 

    Now, I have successfully deployed the model and given a score URI. I can see the deployed model in my workspace and the associated information

    Question: How can I test this deployed model with new cases (remember, I have no python experience that was why i chose the autoML in the first place).

    Thanks

    DrWiy


    • Edited by xplicitLoop Wednesday, August 21, 2019 10:10 AM
    Wednesday, August 21, 2019 10:07 AM

All replies

  • Hi,

    Once you have the best model at hand, it is time to deploy it as a web service to predict on new data.

    Automated ML helps you with deploying the model without writing code:

    1. You have a couple options for deployment.
      • Option 1: To deploy the best model (according to the metric criteria you defined), select Deploy Best Model from the Run Detail page.
      • Option 2: To deploy a specific model iteration from this experiment, drill down on the model to open its run detail page and select Deploy Model.
    2. Populate the Deploy Model pane,
    3. Select Deploy. Deployment can take about 20 minutes to complete.

    The following message appears when deployment successfully completes.

    Now you have an operational web service to generate predictions!

    You can send data to this API (Score URI) and receive the prediction returned by the model. Please refer link below how to create clients for the web service by using C#, Go, Java, and Python.

    scoring_uri is the the REST API address.

    The REST API expects the body of the request to be a JSON document with the following structure:

    {

        "data":

            [

                <model-specific-data-structure>

            ]

    }

    The following document explains how to consume the Score API using different programming languages.

    https://docs.microsoft.com/en-us/azure/machine-learning/service/how-to-consume-web-service#request-data

    Thanks

    Wednesday, August 21, 2019 11:03 AM
    Moderator
  • Thanks.

    Model deployed using AUTOMATED MACHINE LEARNING DIRECTLY FROM WORKSPACE

    https://docs. microsoft.com/en-us/azure/machine-learning/service/tutorial-first-experiment-automated-ml

    Following that tutorial, how can the model deployed be used to test new cases? That is, how can it be consumed as a web service.

    Wednesday, August 21, 2019 12:31 PM
  • Indeed, When you deploy from an Auto ML creates a REST API. The scoring_uri is the REST API address. You can send data to this API and receive the prediction returned by the model. 

    The general workflow for creating a client that uses a machine learning web service is:

    1. Use the SDK to get the connection information.
    2. Determine the type of request data used by the model.
    3. Create an application that calls the web service.





    Wednesday, August 21, 2019 12:39 PM
  • Hi

    Thanks for your explanation.

    Please one more question, I deployed directly to Azure Container Instances (ACI), where authentication is disabled. Can I still go ahead with the request data using C# without using String Key but just using the String URI. Thanks

    Kindly see below:

    Authentication with keys

    When you enable authentication for a deployment, you automatically create authentication keys.

    • Authentication is enabled by default when you are deploying to Azure Kubernetes Service.
    • Authentication is disabled by default when you are deploying to Azure Container Instances.

    To control authentication, use the auth_enabled parameter when you are creating or updating a deployment.

    Thursday, August 22, 2019 8:21 AM
  • Yes, Automated ML Authentication disabled by default for ACI deployment. The auth header Only needed if the web service requires authentication, So you can go ahead with the scoring Uri.

    Thanks


    • Edited by AzureML1256 Friday, August 23, 2019 5:18 AM
    • Proposed as answer by AzureML1256 Friday, August 23, 2019 6:16 AM
    Friday, August 23, 2019 5:16 AM