#optimization #hyper-parameters #black-box-optimization #machine #learning

blackbox

Optimize blackbox functions, suitable for hyperparameter optimization

2 unstable releases

0.2.0 Feb 1, 2020
0.1.0 Mar 2, 2019

#9 in #hyper-parameters


Used in blackbox_derive

MIT license

14KB
265 lines

Blackbox Optimizer

Optimizing the output of any "black box" function. One use case is hyper parameter optimization.

Right now, only random search is implemented but many other optimization methods will be added. Random search and bayesian search is implemented and works well, but there is a lot more to do, including low-hanging fruits such as implementing grid search and other basic search methods.

Dependencies

~4–11MB
~114K SLoC