cotengra.hyperoptimizers.hyper_skopt

Hyper optimization using scikit-optimize.

Classes

SkoptOptLib

Hyper-optimization using scikit-optimize.

Functions

convert_param_to_skopt(param, name)
get_methods_space(methods)
convert_to_skopt_space(method, space)

Module Contents

cotengra.hyperoptimizers.hyper_skopt.convert_param_to_skopt(param, name)

cotengra.hyperoptimizers.hyper_skopt.get_methods_space(methods)

cotengra.hyperoptimizers.hyper_skopt.convert_to_skopt_space(method, space)

class cotengra.hyperoptimizers.hyper_skopt.SkoptOptLib

Bases: cotengra.hyperoptimizers.hyper.HyperOptLib

Hyper-optimization using scikit-optimize.

setup(methods, space, optimizer=None, sampler='et', method_sampler='et', sampler_opts=None, method_sampler_opts=None, **kwargs)

Initialize the scikit-optimize optimizer.

Parameters:

• methods (list[str]) – The list of contraction methods to optimize over.

• space (dict[str, dict[str, dict]]) – The search space.

• optimizer (HyperOptimizer, optional) – The parent optimizer instance.

• sampler (str, optional) –

  The regressor to use to optimize each method’s search space, valid options are:

  • ”et”: Extra Trees Regressor

  • ”rf”: Random Forest Regressor

  • ”gbrt”: Gradient Boosting Regressor

  • ”gp”: Gaussian Process Regressor

• method_sampler (str, optional) – Meta-optimizer to use to select which overall method to use.

• sampler_opts (dict, optional) – Options to supply to the per-method optimizer.

• method_sampler_opts (dict, optional) – Options to supply to the method selector.

get_setting()

Find the next parameters to test.

report_result(setting, trial, score)

Report the result of a trial to the skopt optimizers.