Skip to main content

Model Serving with Kubeflow and Kale

This course will walk you through creating your own Kubeflow deployment using MiniKF, then running a Kubeflow Pipelines workflow with hyperparameter tuning to train and serve a model. You do all that from inside a Jupyter Notebook. You will build a complex data science pipeline with hyperparameter tuning on Kubeflow Pipelines, without using any CLI commands or SDKs. Then, you will easily serve the model and make predictions against new data. You will run the OpenVaccine example, a Kaggle competition related to the need to bring the COVID-19 vaccine to mass production.

  • Course Number

  • Self-Paced

About This Course

Τhis course consists of three main sections. First, you will create a notebook and convert it to a pipeline in Kubeflow Pipelines. This way, you have a time machine for your data and code, and your pipeline will run in the same environment where you have developed your code, without needing to build new Docker images. After running a single pipeline, you will optimize your model using hyperparameter tuning. You are going to use Katib, Kubeflow's official hyperparameter tuner. Kale will orchestrate Katib and KFPexperiments so that every Katib trial is a pipeline run in Kubeflow Pipelines. Ultimatelly, you will take the best model and serve it, using KF Serving, Kubeflow's component for serving models to production.

Frequently Asked Questions

What web browser should I use?

The Open edX platform works best with current versions of Chrome, Edge, Firefox, Internet Explorer, or Safari.

See our list of supported browsers for the most up-to-date information.