To run a sample inference request, complete the following steps:
Get a shell to the client container/pod.
kubectl exec --stdin --tty <<client_pod_name>> -- /bin/bash
Run a sample inference request.
image_client -m resnet50_netdef -s INCEPTION -u <<LoadBalancer_IP_recorded earlier>>:8000 -c 3 images/mug.jpg
This inferencing request calls the
resnet50_netdefmodel that is used for image recognition. Other clients can also send inferencing requests concurrently by following a similar approach and calling out the appropriate model.