In this section, you will setup a web interface that can interact with a trained model server.
The web UI uses a Flask server to host the HTML/CSS/JavaScript files for the web page.
The Python program, mnist_client.py, contains a function that interacts directly with the prediction GRPC API exposed by our model server,
where name mnist-classifier is the deployment name of our model server, address ambassador and port 80 point to the Ambassador endpoint,
which will route the request to the Seldon model service.
The following steps describe how to deploy the docker image in your Kubeflow cluster, the web-ui directory also contains a Dockerfile to build the application into a container image.
Ensure that your model is live and listening for GRPC requests as described in serving.
The folder ks_app contains some pre-defined ksonnet components as we saw in step 3,
including a web-ui component which we will apply now.
cd ks_app
ks apply ${KF_ENV} -c web-uiIn a browser, navigate to the Kubeflow URI https://<name>.endpoints.<project>.cloud.goog/pytorch-ui/,
Ambassador routes requests to the prefix /pytorch-ui/ to our deployed service web-ui
Next: Teardown
Back: Serving the Model