Skip to content

Katib Hyperparameter Tuning

Summary

In this course you have learned to do the following:

  • Evaluate multiple models within a notebook to identify the best model for hyperparamter tuning.
  • Prepare a Jupyter Notebook for hyperparameter tuning by isolating and tagging Pipeline Metrics.
  • Create an execute a Katib Job and review output to identify the ideal hyperparameters.

Next Steps

After identifying your ideal model and ideal hyperparameter you are prepared to finalize and serve your model. This can be done with the KFServing and the SDK from within the notebook - for an example of this please see the public tutorial From Notebook to Kubeflow Pipelines to KFServing: the Data Science Odyssey

Additional Education

Please consider taking our additional courses to expand your continuing Kubeflow education!