cotengra.experimental.hyper_smac

Hyper parameter optimization using SMAC3.

https://automl.github.io/SMAC3/latest/

Classes

SMACOptLib

Hyper-optimization using SMAC3 with per-method facades and

Functions

build_config_space(method, space)	Build a SMAC ConfigurationSpace from a cotengra space dict.
config_to_params(config)	Convert a SMAC Configuration to a plain dict of parameters.

Module Contents

cotengra.experimental.hyper_smac.build_config_space(method, space)

Build a SMAC ConfigurationSpace from a cotengra space dict.

Parameters:

• method (str) – The method name (used as a prefix to avoid name collisions when multiple methods share the same parameter name).

• space (dict[str, dict]) – The search space for a single method.

Returns:

cs

Return type:

ConfigurationSpace

cotengra.experimental.hyper_smac.config_to_params(config)

Convert a SMAC Configuration to a plain dict of parameters.

class cotengra.experimental.hyper_smac.SMACOptLib

Bases: cotengra.experimental.hyper.HyperOptLib

Hyper-optimization using SMAC3 with per-method facades and a Lower Confidence Bound method selector.

setup(methods, space, optimizer=None, facade='BlackBoxFacade', n_trials=10000, seed=0, method_exploration=1.0, method_temperature=1.0, **facade_opts)

get_setting()

report_result(setting, trial, score)