apiVersion: "kubeflow.org/v1alpha1"
kind: StudyJob
metadata:
  namespace: kubeflow
  labels:
    controller-tools.k8s.io: "1.0"
  name: random-example-goal-test
spec:
  studyName: random-example-long-test
  owner: crd
  optimizationtype: maximize
  objectivevaluename: test/accuracy
  optimizationgoal: 0.99
  metricsnames:
    - train/accuracy
    - train/cross_entropy
    - test/accuracy
    - test/cross_entropy
  parameterconfigs:
    - name: --learning_rate
      parametertype: double
      feasible:
        min: "0.001"
        max: "0.01"
    - name: --dropout
      parametertype: double
      feasible:
        min: "0.7"
        max: "0.9"
  workerSpec:
    goTemplate:
        rawTemplate: |-
          apiVersion: batch/v1
          kind: Job
          metadata:
            name: {{.WorkerID}}
            namespace: kubeflow
          spec:
            template:
              spec:
                containers:
                - name: {{.WorkerID}}
                  image: katib/tensorflow-mnist-example:latest
                  command:
                  - "python"
                  - "/mnist_with_summaries.py"
                  - "--log_dir=/log/{{.WorkerID}}"
                  {{- with .HyperParameters}}
                  {{- range .}}
                  - "{{.Name}}={{.Value}}"
                  {{- end}}
                  {{- end}}
                  volumeMounts:
                      - mountPath: "/log"
                        name: "log"
                restartPolicy: Never
                volumes:
                  - name: "log"
                    persistentVolumeClaim:
                        claimName: "tfevent-volume"
  metricsCollectorSpec:
    goTemplate:
      rawTemplate: |-
        apiVersion: batch/v1beta1
        kind: CronJob
        metadata:
          name: {{.WorkerID}}
          namespace: kubeflow
        spec:
          schedule: "*/1 * * * *"
          successfulJobsHistoryLimit: 0
          failedJobsHistoryLimit: 1
          jobTemplate:
            spec:
              template:
                spec:
                  containers:
                  - name: {{.WorkerID}}
                    image: registry.gitlab.com/yujioshima/katib-images/tfevent-metrics-collector
                    args:
                    - "python"
                    - "main.py"
                    - "-m"
                    - "vizier-core"
                    - "-s"
                    - "{{.StudyID}}"
                    - "-w"
                    - "{{.WorkerID}}"
                    - "-d"
                    - "/log/{{.WorkerID}}"
                    volumeMounts:
                        - mountPath: "/log"
                          name: "log"
                  volumes:
                    - name: "log"
                      persistentVolumeClaim:
                          claimName: "tfevent-volume"
                  restartPolicy: Never
                  serviceAccountName: metrics-collector
                  imagePullSecrets:
                      - name: gitlabregcred
  suggestionSpec:
    suggestionAlgorithm: "random"
    requestNumber: 1