In this repository we release all code to replicate all results, tables and figures presented in the paper: A Collection of Quality Diversity Optimization Problems Derived from Hyperparameter Optimization of Machine Learning Models
The repository is structured as follows:
pyribs_ranger.py
and pyribs_xgboost.py
contain code for generating heatmaps on all benchmark problems; they can also be used as an entry point on how to set up YAHPO Gym for our QDO problemsbenchmark_ranger.py
and benchmark_xgboost.py
are used for running the benchmark experimentsrandom_emitter.py
contains code for a RandomEmitter
to be used via pyribshelpers.py
contains helper functionsPipfile
and requirements.txt
list all python module requirementsResults/
contains benchmark results as .csv
filesPlots/
contains all plots as presented in the paperanalysis.R
contains code to analyze benchmarks results and generate fancy ggplot plotsYAHPO Gym v1.0 was used.
Please see here for more information about YAHPO Gym. Detailed documentation on how to setup YAHPO Gym can be found here.