def __init__(
self,
*,
candidates_func: Callable[
[
"torch.Tensor",
"torch.Tensor",
Optional["torch.Tensor"],
"torch.Tensor",
],
"torch.Tensor",
] = None,
constraints_func: Optional[Callable[[FrozenTrial], Sequence[float]]] = None,
n_startup_trials: int = 10,
independent_sampler: Optional[BaseSampler] = None,
):
_imports.check()
self._candidates_func = candidates_func
self._constraints_func = constraints_func
self._independent_sampler = independent_sampler or RandomSampler()
self._n_startup_trials = n_startup_trials
self._study_id: Optional[int] = None
self._search_space = IntersectionSearchSpace()
def test_intersection_search_space_class_with_different_studies() -> None:
search_space = IntersectionSearchSpace()
with StorageSupplier("sqlite") as storage:
study0 = create_study(storage=storage)
study1 = create_study(storage=storage)
search_space.calculate(study0)
with pytest.raises(ValueError):
# An `IntersectionSearchSpace` instance isn't supposed to be used for multiple studies.
search_space.calculate(study1)
def __init__(
self,
consider_prior: bool = True,
prior_weight: float = 1.0,
consider_magic_clip: bool = True,
consider_endpoints: bool = False,
n_startup_trials: int = 10,
n_ei_candidates: int = 24,
gamma: Callable[[int], int] = default_gamma,
weights: Callable[[int], np.ndarray] = default_weights,
seed: Optional[int] = None,
*,
multivariate: bool = False,
warn_independent_sampling: bool = True,
) -> None:
self._parzen_estimator_parameters = _ParzenEstimatorParameters(
consider_prior, prior_weight, consider_magic_clip,
consider_endpoints, weights)
self._prior_weight = prior_weight
self._n_startup_trials = n_startup_trials
self._n_ei_candidates = n_ei_candidates
self._gamma = gamma
self._weights = weights
self._warn_independent_sampling = warn_independent_sampling
self._rng = np.random.RandomState(seed)
self._random_sampler = RandomSampler(seed=seed)
self._multivariate = multivariate
self._search_space = IntersectionSearchSpace()
if multivariate:
warnings.warn(
"``multivariate`` option is an experimental feature."
" The interface can change in the future.",
ExperimentalWarning,
)
def test_intersection_search_space() -> None:
search_space = IntersectionSearchSpace()
study = create_study()
# No trial.
assert search_space.calculate(study) == {}
assert search_space.calculate(study) == intersection_search_space(study)
# First trial.
study.optimize(
lambda t: t.suggest_float("y", -3, 3) + t.suggest_int("x", 0, 10),
n_trials=1)
assert search_space.calculate(study) == {
"x": IntUniformDistribution(low=0, high=10),
"y": UniformDistribution(low=-3, high=3),
}
assert search_space.calculate(study) == intersection_search_space(study)
# Returning sorted `OrderedDict` instead of `dict`.
assert search_space.calculate(study, ordered_dict=True) == OrderedDict([
("x", IntUniformDistribution(low=0, high=10)),
("y", UniformDistribution(low=-3, high=3)),
])
assert search_space.calculate(
study,
ordered_dict=True) == intersection_search_space(study,
ordered_dict=True)
# Second trial (only 'y' parameter is suggested in this trial).
study.optimize(lambda t: t.suggest_float("y", -3, 3), n_trials=1)
assert search_space.calculate(study) == {
"y": UniformDistribution(low=-3, high=3)
}
assert search_space.calculate(study) == intersection_search_space(study)
# Failed or pruned trials are not considered in the calculation of
# an intersection search space.
def objective(trial: Trial, exception: Exception) -> float:
trial.suggest_float("z", 0, 1)
raise exception
study.optimize(lambda t: objective(t, RuntimeError()),
n_trials=1,
catch=(RuntimeError, ))
study.optimize(lambda t: objective(t, TrialPruned()), n_trials=1)
assert search_space.calculate(study) == {
"y": UniformDistribution(low=-3, high=3)
}
assert search_space.calculate(study) == intersection_search_space(study)
# If two parameters have the same name but different distributions,
# those are regarded as different parameters.
study.optimize(lambda t: t.suggest_float("y", -1, 1), n_trials=1)
assert search_space.calculate(study) == {}
assert search_space.calculate(study) == intersection_search_space(study)
# The search space remains empty once it is empty.
study.optimize(
lambda t: t.suggest_float("y", -3, 3) + t.suggest_int("x", 0, 10),
n_trials=1)
assert search_space.calculate(study) == {}
assert search_space.calculate(study) == intersection_search_space(study)