Skip to content

coiled/dask-xgboost-nyctaxi

Repository files navigation

XGBoost + Dask + HPO with Optuna

This repository contains a sequence of notebooks that progressively train a large model on tabular data using ...

  • XGBoost for gradient boosted trees
  • Dask for parallel computing
  • Optuna for hyper-parameter-optimization

We find that this combination is pragmatic in large scale machine learning problems.

Notebooks

The notebooks in this repository are progressively more sophisticated. We start by looking at the data, and a simple training with XGBoost and Dask. Then we show how to do Hyper Parameter Optimization with XGBoost and Dask and Optuna, and finally, we train many models in parallel with XGBoost and Dask and Optuna. This progression can help to make it clear what each tool does, and how best to combine them.

We hope that these notebooks serve as a prototype for others to adapt to their needs.

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Contributors 4

  •  
  •  
  •  
  •