Introducing Prefect-ML: Orchestrate a Distributed Hyperparameter Grid Search on Dask - by Kevin Kho

The post covers

  • How to find the right combination of model hyperparameters and prevent under- and overfitting
  • How to tune hyperparameters in a distributed Dask cluster using Prefect’s mapping
  • Titanic dataset! :passenger_ship:
  • Train test split
  • Training various types of models using scikit-learn incl. logistic regression, random forest, and K-neighbors classifiers
  • How to display the accuracy results for each set of parameters using the Artifact’s API :bar_chart:

Blog post

Code used in the blog