Polyaxon provides native support for several KubeFlow components.

Overview

  1. Kubeflow Operators: Polyaxon can schedule and manage Kubeflow operators natively. Polyaxon provides a uniform workflow for:

    • Viewing logs and resources.
    • Tracking metrics, outputs, and models.
    • Comparing and driving insights. All Kubeflow jobs can be compared and composed natively with other operations supported by Polyaxon.
  2. Kubeflow Pipelines: Polyaxon supports Kubeflow Pipeline components with very few changes.
  3. Kubeflow KFServing: Polyaxon provides reusable components that can deploy models using KFServing.

Deploying Kubeflow operators

For teams not running/using Kubeflow and want to use this integration, Polyaxon provides Helm charts for the Kubeflow operators currently supported.

These Helm charts will be maintained and supported by Polyaxon to allow users to deploy and manage Kubeflow Operators in an easy way.

These operators require Helm to be installed.

We are also distributing these charts directly on our official Helm charts repo https://charts.polyaxon.com

$ helm repo add polyaxon https://charts.polyaxon.com
$ helm repo update

Deploying/Deleting TFJob

In order to use Kubeflow as a backend for running distributed Tensorflow experiments, you need to deploy TFJob on the same namespace where Polyaxon was deployed

helm install plxtf polyaxon/tfjob --namespace=polyaxon
helm install del plxtf --purge

Deploying/Deleting PytorchJob

In order to use Kubeflow as a backend for running distributed Pytorch experiments, you need to deploy PytorchJob on the same namespace where Polyaxon was deployed

helm install plxpytorch polyaxon/pytorchjob --namespace=polyaxon
helm install del plxpytorch --purge

Deploying/Deleting MpiJob

In order to use Kubeflow as a backend for running distributed experiments using MPI, you need to deploy PytorchJob on the same namespace where Polyaxon was deployed

helm install plxmpi polyaxon/mpijob --namespace=polyaxon
helm install del plxmpi --purge