kubeflow #77

Supports: None

KubeFlow Bundle

Overview

This bundle deploys KubeFlow to a Juju k8s model.

KubeFlow consists of:

  • JupyterHub, for running interactive notebooks

  • TensorFlow Job Dashboard, to manage TensorFlow jobs

  • TensorFlow Training, for training TensorFlow models

  • TensorFlow Serving, for serving TensorFlow models

  • Ambassador, an API gateway for managing access to the services

And several charms that will be added soon:

  • Seldon, for deploying ML models

  • PyTorch operator, for deploying PyTorch models

Deploying

Setup

If you are on macOS or Windows, you will need to use an Ubuntu VM. You can install multipass and access an Ubuntu VM with these commands:

multipass launch --name kubeflow --mem 2G
multipass shell kubeflow

Once you have an Ubuntu environment, you'll need to install these snaps to get started:

sudo snap install juju --classic
sudo snap install juju-wait --classic

microk8s

You'll also need to install the microk8s snap:

sudo snap install microk8s --classic

Next, you can run the commands in deploy-microk8s.sh individually, or run the script as a whole.

CDK

You'll also need to install the kubectl snap:

sudo snap install kubectl --classic

You will then need to create an AWS account for juju to use, and then add the credentials to juju:

$ juju add-credential aws
Enter credential name: kubeflow-test

Using auth-type "access-key".

Enter access-key: <YOUR ACCESS KEY>

Enter secret-key: <YOUR SECRET KEY>

Credential "kubeflow-test" added locally for cloud "aws".

Next, you can run the commands in deploy-aws.sh individually, or run the script as a whole.

Using

JupyterHub

JupyterHub is available at /hub/.

TensorFlow Job Dashboard

The TensorFlow Job dashboard is available at /tfjobs/ui/.

TensorFlow Jobs

To submit a TensorFlow job to the dashboard, you can run this kubectl command:

kubectl create -n <NAMESPACE> -f path/to/job/definition.yaml

Where <NAMESPACE> matches the name of the Juju model that you're using, and path/to/job/definition.yaml should point to a TFJob definition similar to the tf_job_mnist.yaml example found here.

TensorFlow Serving

You can submit a model to be served with TensorFlow Serving:

# For a single model
juju deploy cs:~kubeflow-charmers/kubeflow-tf-serving --storage models=storage-class,, --config model=/path/to/base/dir/model-name

# For a model.conf:
juju deploy cs:~kubeflow-charmers/kubeflow-tf-serving --storage models=storage-class,, --config model-conf=/path/to/model.conf

Files

Download .zip

Bundle configuration