How to run in stand-alone mode

[dailong]

How to run in stand-alone mode ，in other words, no ps and worker in v1alpha2 api?

[gaocegege]

You can create just one worker, and use this pod to train locally.

[dailong]

Thank you for your reply. like this? @gaocegege

apiVersion: "kubeflow.org/v1alpha2"
kind: "TFJob"
metadata:
  name: "dist-mnist-sigle"
  namespace: "default"
spec:
  tfReplicaSpecs:
    Worker:
      replicas: 1
      restartPolicy: OnFailure
      template:
        spec:
          containers:
            - name: tensorflow
              image: kubeflow/tf-dist-mnist-test:1.0
              env:
              - name: TEST_TMPDIR
                value: /training
              command: ["python /training/main.py"]
              volumeMounts:
              - name: mnist-data
                mountPath: /training
          volumes:
            - name: mnist-data
              persistentVolumeClaim:
                claimName: mnist-my-a

[gaocegege]

Yeah, and you do not need to write replicas: 1 since the default value for replicas is 1

[gaocegege]

I am going to close the issue. If you have any question please leave comments here.