use-cases/model-fine-tuning-pipeline/data-preparation/gemma-it/manifests/job.yaml

apiVersion: batch/v1
kind: Job
metadata:
  name: data-prep
spec:
  backoffLimit: 0
  template:
    metadata:
      labels:
        app: data-prep
        ml-platform: data-prep
    spec:
      containers:
      - name: job
        image: V_IMAGE_URL
        imagePullPolicy: Always
        env:
        - name: "BUCKET"
          value: "V_DATA_BUCKET"
        - name: "DATASET_INPUT_PATH"
          value: "V_DATASET_INPUT_PATH"
        - name: "DATASET_INPUT_FILE"
          value: "V_DATASET_INPUT_FILE"
        - name: "DATASET_OUTPUT_PATH"
          value: "V_DATASET_OUTPUT_PATH"
        - name: "PROJECT_ID"
          value: "V_PROJECT_ID"
        - name: "PROMPT_MODEL_ID"
          value: "V_PROMPT_MODEL_ID"
        - name: "REGION"
          value: "V_REGION"
        resources:
          requests:
            cpu: 100m
            memory: 512Mi
          limits:
            cpu: 250m
            memory: 1Gi
      nodeSelector:
        resource-type: "cpu"
      restartPolicy: Never
      serviceAccountName: V_KSA
      tolerations:
      - key: "on-demand"
        operator: "Exists"
        effect: "NoSchedule"