Skip to content

Latest commit

 

History

History
24 lines (13 loc) · 1.5 KB

README.md

File metadata and controls

24 lines (13 loc) · 1.5 KB

Scaling Jupyter Notebooks with Kubeflow Pipelines

The KF_MNIST.ipynb notebook is designed to show Kubeflow Pipelines functionality.

You can run this notebook in MiniKF on your local machine. Instructions on how to install Kubeflow on a Kubernetes cluster can be found here. This resource also includes instructions on creating Jupyter Notebooks.

Once Kubeflow is installed, create a Notebook. Once your notebook is ready, you should be able to clone the current repository by opening it and navigating to New -> Terminal.

In the terminal, run the following command:

git clone https://github.com/learnk8s/kubeflow-pipelines-demo.git

Once cloned, if you switch back to the notebook, you will see kubeflow-pipelines-demo folder which will include the KF_MNIST.ipynb notebook.

At the end of the execution of the notebook, in Kubeflow dashboard, navigate to Pipelines -> Experiments and you should see the experiment that notebook just created in the pipelines.

Example Pipelines

If you would like to look at some sample pipelines check out the following repos: