Execute without logs
$ kubectl apply -f <job.yaml>
Execute with logs
$ job-runner.sh <job.yaml>
Observing your GPU job(s)
$ kubectl get jobs -n yn-gpu-workload -l k8s-user=<user_name>
Deleting your GPU job
$ kubectl delete job -n yn-gpu-workload <job_name>