Version: Next

Create with Web UI

To deploy a simple model, we can use the flowers prediction example from KFServing.

  1. To do this, go to Models.
  2. Then click New Model Server.
  3. Copy and paste the yaml below, filling in your namespace for <namespace>, then click Create.
    apiVersion: "serving.kubeflow.org/v1beta1"
    kind: "InferenceService"
    metadata:
    namespace: <namespace>
    name: "flower-sample"
    spec:
    predictor:
    tensorflow:
    storageUri: "gs://kfserving-samples/models/tensorflow/flowers"
  4. Wait for the resource to finish setting up, to confirm it's running you should see a green check as indicated.
  5. To test the model, we can use this JupyterLab notebook