github.com/kubeflow/training-operator@v1.7.0/examples/tensorflow/distribution_strategy/keras-API/multi_worker_tfjob.yaml (about)

     1  apiVersion: kubeflow.org/v1
     2  kind: TFJob
     3  metadata:
     4    name: multi-worker
     5  spec:
     6    runPolicy:
     7      cleanPodPolicy: None
     8    tfReplicaSpecs:
     9      Worker:
    10        replicas: 2
    11        restartPolicy: Never
    12        template:
    13          spec:
    14            containers:
    15              - name: tensorflow
    16                image: kubeflow/tf-multi-worker-strategy:latest
    17                volumeMounts:
    18                  - mountPath: /train
    19                    name: training
    20                resources:
    21                  limits:
    22                    nvidia.com/gpu: 1
    23            volumes:
    24              - name: training
    25                persistentVolumeClaim:
    26                  claimName: strategy-volume