TensorFlow Training for Kubeflow

This charm deploys the TensorFlow Training component of Kubeflow to Kubernetes models in Juju.

Usage

To submit models to be trained, you must create a TFJob custom resource in Kubernetes. For example, to submit the distributed mnist model, which is used for e2e testing, you can use:

kubectl create -n $namespace 
    -f https://raw.githubusercontent.com/kubeflow/tf-operator/master/examples/v1alpha2/dist-mnist/tf_job_mnist.yaml

(Note: The namespace is the name of the Kubernetes model in Juju that this charm is deployed into.)

You can then check the status of the job via either the TensorFlow Dashboard, or kubectl:

kubectl get -o yaml -n $namespace tfjobs dist-mnist-for-e2e-test