smac.optimizer.subspaces.boing_subspace¶
Classes
|
Subspace for BOinG optimizer. |
- class smac.optimizer.subspaces.boing_subspace.BOinGSubspace(config_space, bounds, hps_types, bounds_ss_cont=None, bounds_ss_cat=None, model_local=<class 'smac.epm.gaussian_process.augmented.GloballyAugmentedLocalGaussianProcess'>, model_local_kwargs={}, acq_func_local=<class 'smac.optimizer.acquisition.EI'>, acq_func_local_kwargs=None, rng=None, initial_data=None, activate_dims=None, incumbent_array=None, acq_optimizer_local=None, acq_optimizer_local_kwargs=None)[source]¶
Bases:
smac.optimizer.subspaces.LocalSubspace
Subspace for BOinG optimizer. Each time we create a new epm model for the subspace and optimize to maximize the acquisition function inside this subregion.
- Parameters
acq_optimizer_local (Optional[AcquisitionFunctionMaximizer]) – Subspace optimizer, used to give a set of suggested points. Unlike the optimizer implemented in epm_chooser, this optimizer does not require runhistory objects.
acq_optimizer_local_kwargs (
Optional
[dict
]) – Parameters for acq_optimizer_local