Quick Start

This example demonstrates how to tune the configuration for a Mobilenet model deployed with Tensorflow Serving under Morphling.

For demonstration, we choose two configurations to tune: the first one the CPU cores (resource allocation), and the second one is maximum serving batch size (runtime parameter). We use grid search for configuration sampling.

Submit the configuration tuning experiment

kubectl apply -f https://raw.githubusercontent.com/alibaba/morphling/main/examples/experiment/experiment-mobilenet-grid.yaml

Monitor the tuning experiment status

kubectl get pe
kubectl describe pe

Monitor sampling trials (performance test)

kubectl get trial

Get the searched optimal configuration

kubectl get pe

Expected output:

NAME                        STATE       AGE     OBJECT NAME   OPTIMAL OBJECT VALUE   OPTIMAL PARAMETERS
mobilenet-experiment-grid   Succeeded   5m59s   qps           31                     [map[category:env name:BATCH_SIZE value:2] map[category:resource name:cpu value:2000m] map[category:resource name:memory value:2000Mi]]

Delete the tuning experiment

kubectl delete pe --all