hyperparameters.parameters

hyperparameters.parameters

Classes

Name	Description
ParameterSet	User-friendly interface for defining hyperparameters.

ParameterSet

hyperparameters.parameters.ParameterSet()

User-friendly interface for defining hyperparameters.

This class allows for the definition of a set of hyperparameters including their types, bounds, default values, and transformations. It supports float, integer, and categorical parameters and provides a fluent interface for chaining parameter definitions.

Attributes

Name	Type	Description
_parameters	List[Dict]	List of parameter definitions.
_var_names	List[str]	List of parameter names.
_var_types	List[str]	List of parameter types (‘float’, ‘int’, ‘factor’).
_bounds	List[Union[Tuple, List]]	List of bounds for each parameter.
_defaults	Dict[str, Any]	Dictionary of default values.
_var_trans	List[Optional[str]]	List of variable transformations.

Examples

>>> ps = ParameterSet()
>>> ps.add_float("max_depth", 1, 10, default=3)
ParameterSet(
    max_depth=Parameter(name='max_depth', type='float', check_on_set=True, bounds=(1, 10), default=3),
)

Methods

Name	Description
add_factor	Add a factor (categorical) hyperparameter.
add_float	Add a float hyperparameter.
add_int	Add an integer hyperparameter.
names	Returns a list of parameter names.
sample_default	Returns the default configuration.

add_factor

hyperparameters.parameters.ParameterSet.add_factor(name, choices, default=None)

Add a factor (categorical) hyperparameter.

Parameters

Name	Type	Description	Default
name	str	Name of the parameter.	required
choices	List[str]	List of possible values for the parameter.	required
default	Optional[str]	Default value for the parameter.	None

Returns

Name	Type	Description
ParameterSet	ParameterSet	The instance itself to allow method chaining.

Examples

>>> from spotoptim.hyperparameters import ParameterSet
>>> ps = ParameterSet()
>>> ps.add_factor("optimizer", ["adam", "sgd"], default="adam")

add_float

hyperparameters.parameters.ParameterSet.add_float(
    name,
    low,
    high,
    default=None,
    transform=None,
)

Add a float hyperparameter.

Parameters

Name	Type	Description	Default
name	str	Name of the parameter.	required
low	float	Lower bound of the parameter range.	required
high	float	Upper bound of the parameter range.	required
default	Optional[float]	Default value for the parameter.	None
transform	Optional[str]	Transformation string, e.g., “log”, “log(x)”, “pow(x, 2)”.	None

Returns

Name	Type	Description
ParameterSet	ParameterSet	The instance itself to allow method chaining.

Examples

>>> from spotoptim.hyperparameters import ParameterSet
>>> ps = ParameterSet()
>>> ps.add_float("learning_rate", 0.0001, 0.1, default=0.01, transform="log")

add_int

hyperparameters.parameters.ParameterSet.add_int(
    name,
    low,
    high,
    default=None,
    transform=None,
)

Add an integer hyperparameter.

Parameters

Name	Type	Description	Default
name	str	Name of the parameter.	required
low	int	Lower bound of the parameter range (inclusive).	required
high	int	Upper bound of the parameter range (inclusive).	required
default	Optional[int]	Default value for the parameter.	None
transform	Optional[str]	Transformation string, e.g., “log”, “log(x)”, “pow(x, 2)”.	None

Returns

Name	Type	Description
ParameterSet	ParameterSet	The instance itself to allow method chaining.

Examples

>>> from spotoptim.hyperparameters import ParameterSet
>>> ps = ParameterSet()
>>> ps.add_int("n_estimators", 10, 100, default=50)
>>> ps.add_int("epochs", 2, 5, transform="log")

names

hyperparameters.parameters.ParameterSet.names()

Returns a list of parameter names.

Returns

Name	Type	Description
	List[str]	List[str]: The names of the parameters.

sample_default

hyperparameters.parameters.ParameterSet.sample_default()

Returns the default configuration.

Returns

Name	Type	Description
	Dict[str, Any]	Dict[str, Any]: A dictionary mapping parameter names to their default values.

Examples

>>> ps = ParameterSet()
>>> ps.add_int("x", 1, 10, default=5)
>>> ps.sample_default()
{'x': 5}