Model by Pre-packaged Server

In this tutorial, we will show how to deploy a model by a pre-packaged server. We deploy a IRIS model by SKLearn pre-packaged server.


Enable Model Deployment in Group Management

Remember to enable model deployment in your group, contact your admin if it is not enabled yet.

Tutorial Steps

  1. Go to User Portal and select Deployments.

  2. Then we are in model deployment list page, now clicking on Create Deployment button.

  3. Fill in the Deployment name field with quickstart-iris

    Select the Model Image field with SKLearn server; This is a pre-packaged model server image that can serve scikit-learn model.

    Fill in the Model URI field with gs://seldon-models/sklearn/iris; This path is included the trained model in the Google Cloud Storage.

  4. In the Resources,

    • choose the instance type, here we use the one with configuration (CPU: 0.5 / Memory: 1 G / GPU: 0)

    • leave Replicas as default (1)

  5. Click on Deploy button, then we will be redirected to model deployment list page. Wait for a while and click on Refresh button to check our model is deployed or not.

    When the deployment is deployed successfully, we can click on cell to check its detail.\

  6. We can view some detailed information in detail page, now let's test our deployed model! Copy the endpoint URL and replace the ${YOUR_ENDPOINT_URL} in the following block.

        -H 'Content-Type: application/json' \
        -d '{ "data": {"tensor": {"shape": [1, 4], "values": [5.3, 3.5, 1.4, 0.2]}} }'

    Then copy the entire block to the terminal for execution, and we are sending tensor as request data.

  • Example of request data

    curl -X POST \
        -H 'Content-Type: application/json' \
        -d '{ "data": {"tensor": {"shape": [1, 4], "values": [5.3, 3.5, 1.4, 0.2]}} }'
  • Example of response data (it predicts the species is Iris setosa as the first index has the highest prediction value)

      "data": {
        "names": [
        "tensor": {
          "shape": [
          "values": [
      "meta": {}

Congratulations! We have deployed a model as an endpoint service that can respond requests anytime from everywhere.


Last updated