This topic describes how to deploy a model trained by AutoLearning to Elastic Algorithm Service (EAS) as an online service. Users can consume the service by calling the API of the service over the Internet.

Prerequisites

The model is trained and evaluated. For more information, see Matching recall or Image classification.

Deploy the model

  1. Open the Model trial wizard.
    1. Log on to the Machine Learning Platform for AI console.
    2. In the left-side navigation pane, choose AutoLearning > General Purpose Model Training.
    3. On the AutoLearning page, click Open in the Operation column.
    4. Click the Model trial tab.
  2. In the Model trial wizard, click Go to EAS deployment.
  3. On the Resources And Models page, select a resource type from the Resources Type list. Then, click Next.
    EAS presets the AutoLearning Processor. Therefore, you do not need to manually specify the Processor type.
  4. On the Deployment details and confirmation page, enter a name in the Custom Model Name field.
  5. In the Number Of Instances and Quota fields, click Up arrow or Down arrow to adjust the number of resources.
  6. Click Deploy.
    Visit the Elastic Algorithm Service page. If the status of the model changes to Running in the State column, the model is deployed.

Call the model

  1. On the Elastic Algorithm Service page, click Invoke Intro in the Service Method column.
  2. EAS allows you to call models deployed as online services through public endpoints or Virtual Private Cloud (VPC) endpoints. You can choose a calling method based on your requirements.
    • Call the model through a public endpoint (commonly used method)
      1. On the Invoke Intro page, click the Public Network Invoke tab to view Access address and Token.
      2. Call an API.
    • Call the model through a VPC endpoint
      1. On the Invoke Intro page, click the VPC Invoke tab to view Access address and Token.
      2. Call an API.