Running training and inference on kubernetes clusters
To build and push the container images:
docker build --platform linux/amd64 -t train-image -f Dockerfile-train .
docker tag train-image lvb243/cml:train-image
docker push lvb243/cml:train-image
docker build --platform linux/amd64 -t inference-image -f Dockerfile-inference .
docker tag inference-image lvb243/cml:inference-image
docker push lvb243/cml:inference-image
After creating the cluster, to deploy the training and then inference apps from the gke cluster shell:
kubectl create -f pvc.yaml
kubectl create -f train.yaml
kubectl create -f service.yaml
kubectl create -f inference.yaml