colbert_train.job.yaml

---
apiVersion: "batch/v1"
kind: "Job"
metadata:
  name: "job-colbert-train"
  namespace: "richardmproject"
spec:
  nodeSelector:
    node-role.ida/gputitan: 'true'
  parallelism: 1
  completions: 1
  template:
    metadata:
      labels:
        jobconfig: "job-colber-train"
        app: "colbert"
    spec:
      containers:
      - name: "colbert"
        image: "docker-registry.default.svc:5000/richardmproject/is-colbert:latest"
        resources:
          requests:
            cpu: "1000m"
            memory: "16Gi"
            nvidia.com/gpu: '1'
          limits:
            cpu: "16000m"
            memory: "64Gi"
            nvidia.com/gpu: '1'
        volumeMounts:
        - mountPath: "/nfs/"
          name: "nfs-access"
        imagePullPolicy: "Always"
        workingDir: /nfs/colbert/ColBERT
        command:
        - "/bin/bash"
        args:
        - "trainColbert.sh"
      volumes:
      - name: "nfs-access"
        persistentVolumeClaim:
          claimName: "richardmvol1claim"
      restartPolicy: "OnFailure"
      serviceAccount: "containerroot"