README.md

Introduction

@author: Yingding Wang
@date: 23.02.2024
@version: 1

This workshop introduces the basics regarding Kubeflow Notebook and Kubeflow Pipeline for manifests deployment 1.7.0.

1. Agenda of Kubeflow Workshop (Basics)

Kubeflow Workbench (JupyterLab) Basics

1. Create, Monitor, Restart, Connect (Workbench: Jupyter Notebook)
2. Install, upgrade workbench python packages
3. Share Data Volume in workbenches
4. Git Versioning and Env variable
5. Working with python juypter notebook

Kubeflow Pipeline Basics

1. Create Kubeflow Pipeline from Workbench
2. Kubeflow Pipeline SDK basics
3. Running further examples of Kubeflow Pipeline

2. Further Readings

• Kubeflow features and components: https://www.kubeflow.org/docs/components/
• Canonical Charmed Kubeflow examples: https://github.com/canonical/kubeflow-examples
• Kubeflow examples: https://github.com/kubeflow/examples
• Kubeflow pipeline v2 SDK reference: https://www.kubeflow.org/docs/components/pipelines/v1/sdk/build-pipeline/
• Q&A Kubeflow Community Slack: https://www.kubeflow.org/docs/about/community/

3. Attributions

My best thanks to Andreea Munteanu from Canonical and Ajinkya Bobade from Kubeflow Community for sending me their internal workshop slides and examples. I have incorperated these valuable insights into the content of this workshop with hands-on tutorials.

My deepest thanks to Prof. Dr. Michael Ingrisch and Dr. Balthasar Schachtner for coordinating and facilitating this kubeflow workshop. Without their help, this workshop will not be possible. Also my thanks to Simon Leutner and Thomas Kluge for their support in the Kubeflow system operation and maintenance on-site.

Finally, my thanks to all the Kubeflow maintainers, Kubeflow Community members for creating this fantastic open-source system for making machine learning workflows on Kubernetes simple, portalbe and scalable.