In the following we explain how to use SMAC3. We explain the framework and the options you can use to configure SMAC.
SMAC stands for Sequential Model-Based Algorithm Configuration and uses Bayesian optimization to configure the hyperparameters of an algorithm. Instance features can be used to optimize the algorithm on a certain set of instances.
We provide examples for the usage of SMAC in the Quickstart guide.
SMAC is written in Python 3 and hosted on GitHub.