def get_hyperparameter_search_space(seed) -> ConfigSpaceWrapper:
"""
The Bernoulli NB configuration space based on the search space from
auto-sklearn:
https://github.com/automl/auto-sklearn/blob/master/autosklearn/pipeline/components/classification/bernoulli_nb.py
Parameters
----------
seed: int
Random seed that will be used to sample random configurations
Returns
-------
cs: ConfigSpace.ConfigurationSpace
The configuration space object
"""
cs = ConfigSpace.ConfigurationSpace('sklearn.naive_bayes.BernoulliNB',
seed)
# the smoothing parameter is a non-negative float
# I will limit it to 1000 and put it on a logarithmic scale. (SF)
# Please adjust that, if you know a proper range, this is just a guess.
alpha = ConfigSpace.UniformFloatHyperparameter(name='alpha',
lower=1e-2,
upper=100,
default_value=1,
log=True)
fit_prior = ConfigSpace.CategoricalHyperparameter(name='fit_prior',
choices=[True, False],
default_value=True)
hyperparameters = [alpha, fit_prior]
return ConfigSpaceWrapper(cs, hyperparameters, None)
def prepare_classifier(configuration_space_wrapper: ConfigSpaceWrapper,
task: openml.tasks.OpenMLTask, run_defaults: bool):
configuration_space = configuration_space_wrapper.assemble()
data_name = task.get_dataset().name
data_qualities = task.get_dataset().qualities
data_tuple = (task.task_id, data_name, data_qualities['NumberOfFeatures'],
data_qualities['NumberOfInstances'])
logging.info('Obtained task %d (%s); %s attributes; %s observations' %
data_tuple)
# obtain deserialized classifier
nominal_indices = task.get_dataset().get_features_by_type(
'nominal', [task.target_name])
numeric_indices = task.get_dataset().get_features_by_type(
'numeric', [task.target_name])
if configuration_space_wrapper.wrapped_in_pipeline:
classifier = sklearnbot.sklearn.as_pipeline(configuration_space,
numeric_indices,
nominal_indices)
else:
classifier = sklearnbot.sklearn.as_estimator(configuration_space,
False)
# sample configuration and set hyperparameters
if not run_defaults:
configuration = configuration_space.sample_configuration(1)
logging.info('Configuration: %s' % configuration.get_dictionary())
classifier.set_params(**configuration.get_dictionary())
else:
logging.info('Running default configuration')
return classifier
def get_hyperparameter_search_space(seed) -> ConfigSpaceWrapper:
"""
The SVM configuration space based on the search space from
auto-sklearn:
https://github.com/automl/auto-sklearn/blob/master/autosklearn/pipeline/components/classification/libsvm_svc.py
Parameters
----------
seed: int
Random seed that will be used to sample random configurations
Returns
-------
cs: ConfigSpace.ConfigurationSpace
The configuration space object
"""
cs = ConfigSpace.ConfigurationSpace('sklearn.svm.SVC', seed)
C = ConfigSpace.UniformFloatHyperparameter(name='C',
lower=0.03125,
upper=32768,
log=True,
default_value=1.0)
kernel = ConfigSpace.CategoricalHyperparameter(
name='kernel', choices=['rbf', 'poly', 'sigmoid'], default_value='rbf')
degree = ConfigSpace.UniformIntegerHyperparameter(name='degree',
lower=1,
upper=5,
default_value=3)
gamma = ConfigSpace.UniformFloatHyperparameter(name='gamma',
lower=3.0517578125e-05,
upper=8,
log=True,
default_value=0.1)
coef0 = ConfigSpace.UniformFloatHyperparameter(name='coef0',
lower=-1,
upper=1,
default_value=0)
shrinking = ConfigSpace.CategoricalHyperparameter(name='shrinking',
choices=[True, False],
default_value=True)
tol = ConfigSpace.UniformFloatHyperparameter(name='tol',
lower=1e-5,
upper=1e-1,
default_value=1e-3,
log=True)
max_iter = ConfigSpace.UnParametrizedHyperparameter('max_iter', -1)
hyperparameters = [
C, kernel, degree, gamma, coef0, shrinking, tol, max_iter
]
degree_depends_on_poly = ConfigSpace.EqualsCondition(
degree, kernel, 'poly')
coef0_condition = ConfigSpace.InCondition(coef0, kernel,
['poly', 'sigmoid'])
conditions = [degree_depends_on_poly, coef0_condition]
return ConfigSpaceWrapper(cs, hyperparameters, conditions)
def get_hyperparameter_search_space(seed) -> ConfigSpaceWrapper:
"""
k-NN search space based on a best effort using the scikit-learn
implementation.
Parameters
----------
seed: int
Random seed that will be used to sample random configurations
Returns
-------
cs: ConfigSpace.ConfigurationSpace
The configuration space object
"""
cs = ConfigSpace.ConfigurationSpace(
'sklearn.neighbors.KNeighborsClassifier', seed)
n_neighbors = ConfigSpace.hyperparameters.UniformIntegerHyperparameter(
name='n_neighbors', lower=1, upper=20, default_value=5)
weights = ConfigSpace.hyperparameters.CategoricalHyperparameter(
name='weights',
choices=['uniform', 'distance'],
default_value='uniform')
algorithm = ConfigSpace.hyperparameters.CategoricalHyperparameter(
name='algorithm',
choices=['auto', 'ball_tree', 'kd_tree', 'brute'],
default_value='auto')
leaf_size = ConfigSpace.UniformIntegerHyperparameter(name='leaf_size',
lower=1,
upper=50,
default_value=1)
p = ConfigSpace.UniformIntegerHyperparameter(name='p',
lower=1,
upper=5,
default_value=2)
metric = ConfigSpace.hyperparameters.CategoricalHyperparameter(
name='metric',
choices=[
'euclidean', 'manhattan', 'chebyshev', 'minkowski', 'wminkowski',
'seuclidean', 'mahalanobis'
],
default_value='minkowski')
hyperparameters = [
n_neighbors,
weights,
algorithm,
leaf_size,
p,
metric,
]
leaf_size_condition = ConfigSpace.InCondition(leaf_size, algorithm,
['ball_tree', 'kd_tree'])
conditions = [leaf_size_condition]
return ConfigSpaceWrapper(cs, hyperparameters, conditions)
def get_hyperparameter_search_space(seed) -> ConfigSpaceWrapper:
"""
The random forest configuration space based on the search space from
auto-sklearn:
https://github.com/automl/auto-sklearn/blob/master/autosklearn/pipeline/components/classification/random_forest.py
Parameters
----------
seed: int
Random seed that will be used to sample random configurations
Returns
-------
cs: ConfigSpace.ConfigurationSpace
The configuration space object
"""
cs = ConfigSpace.ConfigurationSpace(
'sklearn.ensemble.RandomForestClassifier', seed)
# TODO: parameterize the number of estimators?
n_estimators = ConfigSpace.Constant(name='n_estimators', value=100)
criterion = ConfigSpace.CategoricalHyperparameter(
name='criterion', choices=['gini', 'entropy'], default_value='gini')
max_features = ConfigSpace.UniformFloatHyperparameter(name='max_features',
lower=0.,
upper=1.,
default_value=0.5)
# max_depth = ConfigSpace.UnParametrizedHyperparameter(
# name='randomforestclassifier__max_depth', value=None)
min_samples_split = ConfigSpace.UniformIntegerHyperparameter(
name='min_samples_split', lower=2, upper=20, default_value=2)
min_samples_leaf = ConfigSpace.UniformIntegerHyperparameter(
name='min_samples_leaf', lower=1, upper=20, default_value=1)
min_weight_fraction_leaf = ConfigSpace.UnParametrizedHyperparameter(
name='min_weight_fraction_leaf', value=0.)
# max_leaf_nodes = ConfigSpace.UnParametrizedHyperparameter(
# name='randomforestclassifier__max_leaf_nodes', value=None)
bootstrap = ConfigSpace.CategoricalHyperparameter(name='bootstrap',
choices=[True, False],
default_value=True)
hyperparameters = [
n_estimators,
criterion,
max_features,
# max_depth,
min_samples_split,
min_samples_leaf,
min_weight_fraction_leaf,
# max_leaf_nodes,
bootstrap
]
return ConfigSpaceWrapper(cs, hyperparameters, None)
def run_optimizer_on_task(task_id: int,
configuration_space_wrapper: ConfigSpaceWrapper,
output_dir: str,
upload_and_delete: bool) \
-> typing.Tuple[bool, typing.Optional[int], typing.Optional[str]]:
local_run_dir = os.path.join(output_dir, str(task_id), str(uuid.uuid4()))
try:
# obtain task
task = openml.tasks.get_task(task_id)
configuration_space = configuration_space_wrapper.assemble()
data_name = task.get_dataset().name
data_qualities = task.get_dataset().qualities
data_tuple = (task.task_id, data_name,
data_qualities['NumberOfFeatures'],
data_qualities['NumberOfInstances'])
logging.info('Obtained task %d (%s); %s attributes; %s observations' %
data_tuple)
# obtain deserialized classifier
nominal_indices = task.get_dataset().get_features_by_type(
'nominal', [task.target_name])
numeric_indices = task.get_dataset().get_features_by_type(
'numeric', [task.target_name])
# obtain prepared classifier
optimizer = sklearnbot.sklearn.as_search_cv(configuration_space,
nominal_indices,
numeric_indices)
# invoke OpenML run
run = openml.runs.run_model_on_task(optimizer, task)
score = run.get_metric_fn(sklearn.metrics.accuracy_score)
logging.info('Task %d - %s; Accuracy: %0.2f' %
(task_id, task.get_dataset().name, score.mean()))
run.to_filesystem(local_run_dir, store_model=False)
if upload_and_delete:
run = run.publish()
shutil.rmtree(local_run_dir)
local_run_dir = None
return True, run.run_id, local_run_dir
except openml.exceptions.OpenMLServerException:
traceback.print_exc()
return False, None, local_run_dir
except openml.exceptions.OpenMLRunsExistError:
traceback.print_exc()
return False, None, local_run_dir
def get_hyperparameter_search_space(seed) -> ConfigSpaceWrapper:
"""
The adaboost configuration space based on the search space from
auto-sklearn:
https://github.com/automl/auto-sklearn/blob/master/autosklearn/pipeline/components/classification/adaboost.py
Parameters
----------
seed: int
Random seed that will be used to sample random configurations
Returns
-------
cs: ConfigSpace.ConfigurationSpace
The configuration space object
"""
cs = ConfigSpace.ConfigurationSpace(
'sklearn.ensemble.AdaBoostClassifier',
seed,
meta={"base_estimator": DecisionTreeClassifier(random_state=0)})
n_estimators = ConfigSpace.hyperparameters.UniformIntegerHyperparameter(
name="n_estimators", lower=50, upper=500, default_value=50, log=False)
learning_rate = ConfigSpace.hyperparameters.UniformFloatHyperparameter(
name="learning_rate", lower=0.01, upper=2, default_value=0.1, log=True)
algorithm = ConfigSpace.hyperparameters.CategoricalHyperparameter(
name="algorithm",
choices=["SAMME.R", "SAMME"],
default_value="SAMME.R")
max_depth = ConfigSpace.hyperparameters.UniformIntegerHyperparameter(
name="base_estimator__max_depth",
lower=1,
upper=10,
default_value=1,
log=False)
hyperparameters = [n_estimators, learning_rate, algorithm, max_depth]
return ConfigSpaceWrapper(cs, hyperparameters, None)
def get_hyperparameter_search_space(seed: int) -> ConfigSpaceWrapper:
"""
The decision tree configuration space based on the search space from
auto-sklearn:
https://github.com/automl/auto-sklearn/blob/master/autosklearn/pipeline/components/classification/decision_tree.py
Parameters
----------
seed: int
Random seed that will be used to sample random configurations
Returns
-------
cs: ConfigSpace.ConfigurationSpace
The configuration space object
"""
cs = ConfigSpace.ConfigurationSpace("sklearn.tree.DecisionTreeClassifier",
seed)
criterion = ConfigSpace.CategoricalHyperparameter(
name='criterion', choices=['gini', 'entropy'], default_value='gini')
min_samples_split = ConfigSpace.UniformIntegerHyperparameter(
name='min_samples_split', lower=2, upper=20, default_value=2)
min_samples_leaf = ConfigSpace.UniformIntegerHyperparameter(
name='min_samples_leaf', lower=1, upper=20, default_value=1)
min_weight_fraction_leaf = ConfigSpace.Constant(
name='min_weight_fraction_leaf', value=0.0)
max_features = ConfigSpace.UnParametrizedHyperparameter(
name='max_features', value=1.0)
min_impurity_decrease = ConfigSpace.UnParametrizedHyperparameter(
'min_impurity_decrease', 0.0)
# TODO: max_leaf_nodes one can only be tuned once config space allows for this.
hyperparameters = [
criterion, max_features, min_samples_split, min_samples_leaf,
min_weight_fraction_leaf, min_impurity_decrease
]
return ConfigSpaceWrapper(cs, hyperparameters, None)
def get_hyperparameter_search_space(seed) -> ConfigSpaceWrapper:
"""
Gradient Boosting search space based on a best effort using the scikit-learn
implementation. Note that for state of the art performance, other packages,
such as xgboost, could be preferred.
Parameters
----------
seed: int
Random seed that will be used to sample random configurations
Returns
-------
cs: ConfigSpace.ConfigurationSpace
The configuration space object
"""
cs = ConfigSpace.ConfigurationSpace(
'sklearn.ensemble.GradientBoostingClassifier', seed)
# fixed to deviance, as exponential requires two classes
loss = ConfigSpace.hyperparameters.Constant(
name='gradientboostingclassifier__loss', value='deviance')
# JvR: changed after conversation with AM on 2019-01-17
learning_rate = ConfigSpace.hyperparameters.UniformFloatHyperparameter(
name='learning_rate',
lower=0.00001,
upper=0.1,
default_value=0.0001,
log=True)
n_estimators = ConfigSpace.hyperparameters.UniformIntegerHyperparameter(
name='n_estimators', lower=64, upper=2048, default_value=100, log=True)
subsample = ConfigSpace.UniformFloatHyperparameter(name='subsample',
lower=0.0,
upper=1.0,
default_value=1.0)
criterion = ConfigSpace.hyperparameters.CategoricalHyperparameter(
name='criterion', choices=['friedman_mse', 'mse', 'mae'])
min_samples_split = ConfigSpace.hyperparameters.UniformIntegerHyperparameter(
name='min_samples_split', lower=2, upper=20, default_value=2)
min_samples_leaf = ConfigSpace.hyperparameters.UniformIntegerHyperparameter(
name='min_samples_leaf', lower=1, upper=20, default_value=1)
# TODO: upper bound?
min_weight_fraction_leaf = ConfigSpace.hyperparameters.UniformFloatHyperparameter(
name='min_weight_fraction_leaf',
lower=0.0,
upper=0.5,
default_value=0.0)
# JvR: changed after conversation with AM on 2019-01-17
max_depth = ConfigSpace.hyperparameters.UniformIntegerHyperparameter(
name='max_depth', lower=1, upper=32, default_value=3)
# TODO: upper bound?
min_impurity_decrease = ConfigSpace.hyperparameters.UniformFloatHyperparameter(
name='min_impurity_decrease', lower=0.0, upper=1.0, default_value=0.0)
max_features = ConfigSpace.hyperparameters.UniformFloatHyperparameter(
name='max_features', lower=0.0, upper=1.0, default_value=0.0)
validation_fraction = ConfigSpace.UniformFloatHyperparameter(
name='validation_fraction', lower=0, upper=1, default_value=0.1)
n_iter_no_change = ConfigSpace.UniformIntegerHyperparameter(
name='n_iter_no_change', lower=1, upper=2048, default_value=200)
tol = ConfigSpace.UniformFloatHyperparameter(name='tol',
lower=1e-5,
upper=1e-1,
default_value=1e-4,
log=True)
hyperparameters = [
loss,
learning_rate,
n_estimators,
subsample,
criterion,
min_samples_split,
min_samples_leaf,
min_weight_fraction_leaf,
max_depth,
min_impurity_decrease,
max_features,
validation_fraction,
n_iter_no_change,
tol,
]
return ConfigSpaceWrapper(cs, hyperparameters, None)
def get_hyperparameter_search_space(seed) -> ConfigSpaceWrapper:
"""
The random forest configuration space based on the search space from
auto-sklearn:
https://github.com/automl/auto-sklearn/blob/master/autosklearn/pipeline/components/classification/sgd.py
Parameters
----------
seed: int
Random seed that will be used to sample random configurations
Returns
-------
cs: ConfigSpace.ConfigurationSpace
The configuration space object
"""
cs = ConfigSpace.ConfigurationSpace('sklearn.linear_model.SGDClassifier',
seed)
loss = ConfigSpace.CategoricalHyperparameter(name='loss',
choices=[
'hinge', 'log',
'modified_huber',
'squared_hinge',
'perceptron'
],
default_value='log')
penalty = ConfigSpace.CategoricalHyperparameter(
name='penalty', choices=['l1', 'l2', 'elasticnet'], default_value='l2')
alpha = ConfigSpace.UniformFloatHyperparameter(name='alpha',
lower=1e-7,
upper=1e-1,
log=True,
default_value=0.0001)
l1_ratio = ConfigSpace.UniformFloatHyperparameter(name='l1_ratio',
lower=1e-9,
upper=1,
log=True,
default_value=0.15)
# fit_intercept = ConfigSpace.UnParametrizedHyperparameter(name='fit_intercept', value=True)
tol = ConfigSpace.UniformFloatHyperparameter(name='tol',
lower=1e-5,
upper=1e-1,
log=True,
default_value=1e-4)
epsilon = ConfigSpace.UniformFloatHyperparameter(name='epsilon',
lower=1e-5,
upper=1e-1,
default_value=1e-4,
log=True)
learning_rate = ConfigSpace.CategoricalHyperparameter(
name='learning_rate',
choices=['optimal', 'invscaling', 'constant'],
default_value='invscaling')
eta0 = ConfigSpace.UniformFloatHyperparameter(name='eta0',
lower=1e-7,
upper=1e-1,
default_value=0.01,
log=True)
power_t = ConfigSpace.UniformFloatHyperparameter('power_t',
1e-5,
1,
default_value=0.5)
average = ConfigSpace.CategoricalHyperparameter(name='average',
choices=[False, True],
default_value=False)
hyperparameters = [
loss,
penalty,
alpha,
l1_ratio,
# fit_intercept,
tol,
epsilon,
learning_rate,
eta0,
power_t,
average
]
# TODO MF: add passive/aggressive here, although not properly documented?
elasticnet = ConfigSpace.EqualsCondition(l1_ratio, penalty, 'elasticnet')
epsilon_condition = ConfigSpace.EqualsCondition(epsilon, loss,
'modified_huber')
power_t_condition = ConfigSpace.EqualsCondition(power_t, learning_rate,
'invscaling')
# eta0 is only relevant if learning_rate!='optimal' according to code
# https://github.com/scikit-learn/scikit-learn/blob/0.19.X/sklearn/
# linear_model/sgd_fast.pyx#L603
eta0_in_inv_con = ConfigSpace.InCondition(eta0, learning_rate,
['invscaling', 'constant'])
conditions = [
elasticnet, epsilon_condition, power_t_condition, eta0_in_inv_con
]
return ConfigSpaceWrapper(cs, hyperparameters, conditions)
def get_hyperparameter_search_space(seed: int) -> ConfigSpaceWrapper:
"""
The extra trees configuration space based on the search space from
auto-sklearn:
https://github.com/automl/auto-sklearn/blob/master/autosklearn/pipeline/components/classification/extra_trees.py
Parameters
----------
seed: int
Random seed that will be used to sample random configurations
Returns
-------
cs: ConfigSpace.ConfigurationSpace
The configuration space object
"""
cs = ConfigSpace.ConfigurationSpace(
'sklearn.ensemble.ExtraTreesClassifier', seed)
# TODO: parameterize the number of estimators?
n_estimators = ConfigSpace.Constant(
name='extratreesclassifier__n_estimators', value=100)
criterion = ConfigSpace.CategoricalHyperparameter(
name='criterion', choices=['gini', 'entropy'], default_value='gini')
# The maximum number of features used in the forest is calculated as m^max_features, where
# m is the total number of features, and max_features is the hyperparameter specified below.
# The default is 0.5, which yields sqrt(m) features as max_features in the estimator. This
# corresponds with Geurts' heuristic.
max_features = ConfigSpace.UniformFloatHyperparameter(name='max_features',
lower=0.,
upper=1.,
default_value=0.5)
# max_depth = ConfigSpace.UnParametrizedHyperparameter(name='extratreesclassifier__max_depth', value=None)
min_samples_split = ConfigSpace.UniformIntegerHyperparameter(
name='min_samples_split', lower=2, upper=20, default_value=2)
min_samples_leaf = ConfigSpace.UniformIntegerHyperparameter(
name='min_samples_leaf', lower=1, upper=20, default_value=1)
min_weight_fraction_leaf = ConfigSpace.UnParametrizedHyperparameter(
name='min_weight_fraction_leaf', value=0.)
# max_leaf_nodes = ConfigSpace.UnParametrizedHyperparameter(name='max_leaf_nodes', value=None)
min_impurity_decrease = ConfigSpace.UnParametrizedHyperparameter(
name='min_impurity_decrease', value=0.0)
bootstrap = ConfigSpace.CategoricalHyperparameter('bootstrap',
[True, False],
default_value=False)
hyperparameters = [
n_estimators,
criterion,
max_features,
# max_depth,
min_samples_split,
min_samples_leaf,
min_weight_fraction_leaf,
# max_leaf_nodes,
min_impurity_decrease,
bootstrap
]
return ConfigSpaceWrapper(cs, hyperparameters, None)
def get_hyperparameter_search_space(seed) -> ConfigSpaceWrapper:
"""
Neural Network search space based on a best effort using the scikit-learn
implementation. Note that for state of the art performance, other packages
could be preferred.
Parameters
----------
seed: int
Random seed that will be used to sample random configurations
Returns
-------
cs: ConfigSpace.ConfigurationSpace
The configuration space object
"""
cs = ConfigSpace.ConfigurationSpace('sklearn.neural_network.MLPClassifier',
seed)
hidden_layer_sizes = ConfigSpace.UniformIntegerHyperparameter(
name='hidden_layer_sizes', lower=32, upper=2048, default_value=2048)
activation = ConfigSpace.CategoricalHyperparameter(
name='activation',
choices=['identity', 'logistic', 'tanh', 'relu'],
default_value='relu')
solver = ConfigSpace.CategoricalHyperparameter(
name='solver', choices=['lbfgs', 'sgd', 'adam'], default_value='adam')
alpha = ConfigSpace.UniformFloatHyperparameter(name='alpha',
lower=1e-5,
upper=1e-1,
log=True,
default_value=1e-4)
batch_size = ConfigSpace.UniformIntegerHyperparameter(name='batch_size',
lower=32,
upper=4096,
default_value=200)
learning_rate = ConfigSpace.CategoricalHyperparameter(
name='learning_rate',
choices=['constant', 'invscaling', 'adaptive'],
default_value='constant')
learning_rate_init = ConfigSpace.UniformFloatHyperparameter(
name='learning_rate_init',
lower=1e-5,
upper=1e-1,
log=True,
default_value=1e-04)
# TODO: Sensible range??
power_t = ConfigSpace.UniformFloatHyperparameter(name='power_t',
lower=1e-5,
upper=1,
log=True,
default_value=0.5)
max_iter = ConfigSpace.UniformIntegerHyperparameter(name='max_iter',
lower=64,
upper=1024,
default_value=200)
shuffle = ConfigSpace.CategoricalHyperparameter(name='shuffle',
choices=[True, False],
default_value=True)
tol = ConfigSpace.UniformFloatHyperparameter(name='tol',
lower=1e-5,
upper=1e-1,
default_value=1e-4,
log=True)
# TODO: log-scale?
momentum = ConfigSpace.UniformFloatHyperparameter(name='momentum',
lower=0,
upper=1,
default_value=0.9)
nesterovs_momentum = ConfigSpace.CategoricalHyperparameter(
name='nesterovs_momentum', choices=[True, False], default_value=True)
early_stopping = ConfigSpace.CategoricalHyperparameter(
name='early_stopping', choices=[True, False], default_value=True)
validation_fraction = ConfigSpace.UniformFloatHyperparameter(
name='validation_fraction', lower=0, upper=1, default_value=0.1)
beta_1 = ConfigSpace.UniformFloatHyperparameter(name='beta_1',
lower=0,
upper=1,
default_value=0.9)
beta_2 = ConfigSpace.UniformFloatHyperparameter(name='beta_2',
lower=0,
upper=1,
default_value=0.999)
n_iter_no_change = ConfigSpace.UniformIntegerHyperparameter(
name='n_iter_no_change', lower=1, upper=1024, default_value=200)
hyperparameters = [
hidden_layer_sizes,
activation,
solver,
alpha,
batch_size,
learning_rate,
learning_rate_init,
power_t,
max_iter,
shuffle,
tol,
momentum,
nesterovs_momentum,
early_stopping,
validation_fraction,
beta_1,
beta_2,
n_iter_no_change,
]
batch_size_condition = ConfigSpace.InCondition(batch_size, solver,
['sgd', 'adam'])
learning_rate_init_condition = ConfigSpace.InCondition(
learning_rate_init, solver, ['sgd', 'adam'])
power_t_condition = ConfigSpace.EqualsCondition(power_t, solver, 'sgd')
shuffle_confition = ConfigSpace.InCondition(shuffle, solver,
['sgd', 'adam'])
tol_condition = ConfigSpace.InCondition(tol, learning_rate,
['constant', 'invscaling'])
momentum_confition = ConfigSpace.EqualsCondition(momentum, solver, 'sgd')
nesterovs_momentum_confition_solver = ConfigSpace.EqualsCondition(
nesterovs_momentum, solver, 'sgd')
nesterovs_momentum_confition_momentum = ConfigSpace.GreaterThanCondition(
nesterovs_momentum, momentum, 0)
nesterovs_momentum_conjunstion = ConfigSpace.AndConjunction(
nesterovs_momentum_confition_solver,
nesterovs_momentum_confition_momentum)
early_stopping_condition = ConfigSpace.InCondition(early_stopping, solver,
['sgd', 'adam'])
validation_fraction_condition = ConfigSpace.EqualsCondition(
validation_fraction, early_stopping, True)
beta_1_condition = ConfigSpace.EqualsCondition(beta_1, solver, 'adam')
beta_2_condition = ConfigSpace.EqualsCondition(beta_2, solver, 'adam')
n_iter_no_change_condition_solver = ConfigSpace.InCondition(
n_iter_no_change, solver, ['sgd', 'adam'])
conditions = [
batch_size_condition, learning_rate_init_condition, power_t_condition,
shuffle_confition, tol_condition, momentum_confition,
nesterovs_momentum_conjunstion, early_stopping_condition,
validation_fraction_condition, beta_1_condition, beta_2_condition,
n_iter_no_change_condition_solver
]
return ConfigSpaceWrapper(cs, hyperparameters, conditions)