deephyper.skopt.utils

deephyper.skopt.utils#

Submodule of utility functions for skopt.

Functions

check_dimension_names

Check whether all dimensions have names.

check_list_types

Check whether all elements of a list x are of the correct type(s).

check_x_in_space

cook_estimator

Cook a default estimator.

cook_initial_point_generator

Cook a default initial point generator.

cook_objective_scaler

Prepare a Scikit-Learn preprocessing pipeline to transform the objective.

create_result

Initialize an OptimizeResult object.

deepcopy

Deep copy operation on arbitrary Python objects.

dimensions_aslist

Convert a dict representation of a search space into a list of dimensions.

dump

Store an skopt optimization result into a file.

eval_callbacks

Evaluate list of callbacks on result.

expected_minimum

Compute the minimum over the predictions of the last surrogate model.

expected_minimum_random_sampling

Minimum search by doing naive random sampling.

has_gradients

Check if an estimator's predict method provides gradients.

is_2Dlistlike

is_listlike

load

Reconstruct a skopt optimization result from a file persisted with deephyper.skopt.dump.

normalize_dimensions

Create a Space where all dimensions are normalized to unit range.

point_asdict

Convert a list representation of a point from a search space to a dictionary representation.

point_aslist

Convert a dictionary representation of a point from a search space to a list representation.

use_named_args

Decorator.

wraps

Decorator factory to apply update_wrapper() to a wrapper function

Classes

ConstantKernel

Dimension

Base class for search space dimensions.

ExtraTreesRegressor

ExtraTreesRegressor that supports conditional standard deviation.

GaussianProcessRegressor

GaussianProcessRegressor that allows noise tunability.

GradientBoostingQuantileRegressor

Predict several quantiles with one estimator.

Grid

Generate samples from a regular grid.

Halton

Creates Halton sequence samples.

Hammersly

Creates Hammersley sequence samples.

HammingKernel

The HammingKernel is used to handle categorical inputs.

InitialPointGenerator

Lhs

Latin hypercube sampling

Matern

OrderedDict

Dictionary that remembers insertion order

RandomForestRegressor

RandomForestRegressor that supports conditional std computation.

Sobol

Generates a new quasirandom Sobol' vector with each call.

Space

Initialize a search space from given specifications.