Enum OptIn 

pub enum OptIn {
    LogScoring,
    Multivariate,
    DiscreteKernels,
    LinearSplit,
    OptunaBandwidth,
    Sharpening,
    Scramble,
    Halton,
    SchemaOrder,
    Ard,
    OptunaSearch,
    MixedSpace,
    Active,
    Separable,
    OptunaInit,
    Margin,
    GpCostBudget,
}

One sampler setting an external benchmark arm can switch on or off.

Most settings are off by default in the sampler itself (plan decisions D105, D107, D108, D110, D111); the promoted ones (D117) are on by default. SamplerKind::build_for builds the sampler’s own default; a SamplerConfig carries departures from it explicitly, so atune-solver --opt-in and --opt-out can put each configuration next to its Optuna counterpart. Only the single-objective, sequential settings are here: kurobako studies are single-objective and the adapter does not advertise CONCURRENT, so TPE’s HSSP split and GP-EI’s pending, constrained and multi-objective acquisitions would be inert, and CMA-ES warm start has no source study to start from.

Variants

LogScoring

[Tpe::with_log_scoring]: mixtures summed with logsumexp.

Multivariate

[Tpe::with_multivariate]: one mixture of product kernels per group (Optuna TPESampler(multivariate=True)).

DiscreteKernels

[Tpe::with_discrete_kernels]: grid-cell mass for integer and stepped values, Aitchison–Aitken kernels for categoricals.

LinearSplit

[Tpe::with_linear_split]: the good set is min(ceil(0.1·n), 25) (Optuna TPESampler’s default gamma).

OptunaBandwidth

[Tpe::with_optuna_bandwidth]: Optuna’s per-estimator numeric bandwidths (gaps among each estimator’s observations, magic clip).

Sharpening

[Tpe::with_sharpening]: twice the candidates, half from a sharpened good-set mixture, scored by the sharpened ratio (atune’s own).

Scramble

[Qmc::with_scramble], seeded by the study seed (Optuna QMCSampler(scramble=True, seed=seed)).

Halton

[Qmc::new_with_halton]: the Halton sequence instead of Sobol (Optuna QMCSampler(qmc_type="halton")).

SchemaOrder

[Qmc::with_schema_order]: assign QMC dimensions by declaration order.

Ard

[GpEi::with_ard]: one Matérn-5/2 lengthscale per dimension, by MAP.

OptunaSearch

[GpEi::with_optuna_search]: Optuna-style startup and acquisition search.

MixedSpace

[GpEi::with_mixed_space]: Hamming distance on categorical axes and exhaustive refinement of small discrete axes.

Active

[Cmaes::with_active]: negative recombination weights.

Separable

[Cmaes::with_separable]: sep-CMA-ES (Optuna use_separable_cma=True).

OptunaInit

[Cmaes::with_optuna_init]: Optuna’s startup, sigma, and SepCMA resampling.

Margin

[Cmaes::with_margin]: CMA-ES with Margin (Optuna with_margin=True).

GpCostBudget

[AutoSamplerSettings::with_gp_cost_budget]: route auto’s GP-EI pick to TPE once the per-ask cost estimate exceeds the budget ([DEFAULT_GP_COST_BUDGET] unless an explicit value is given).

Implementations

impl OptIn

pub const ALL: [Self; 17]

Every setting, grouped by the sampler kind it applies to.

pub const fn label(self) -> &'static str

The flag label (log-scoring, multivariate, …), named after the sampler’s builder method.

pub const fn kind(self) -> SamplerKind

The sampler kind this setting belongs to.

pub const fn is_default(self) -> bool

Whether the sampler turns this setting on by default (D117): the six TPE parity settings, QMC schema order, GP-EI’s Optuna search, and CMA-ES Optuna initialization. Only these accept --opt-out, and an --opt-in of one is a no-op.

Trait Implementations

impl Clone for OptIn

fn clone(&self) -> OptIn

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Copy for OptIn

impl Debug for OptIn

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl Eq for OptIn

impl FromStr for OptIn

Parses a flag label; the error is a message listing every accepted label.

type Err = String

The associated error which can be returned from parsing.

fn from_str(label: &str) -> Result<Self, String>

Parses a string s to return a value of this type.

impl Hash for OptIn

fn hash<__H: Hasher>(&self, state: &mut __H)

Feeds this value into the given Hasher.

fn hash_slice<H>(data: &[Self], state: &mut H)

where H: Hasher, Self: Sized,

Feeds a slice of this type into the given Hasher.

impl Ord for OptIn

fn cmp(&self, other: &OptIn) -> Ordering

This method returns an Ordering between self and other.

fn max(self, other: Self) -> Self

where Self: Sized,

Compares and returns the maximum of two values.

fn min(self, other: Self) -> Self

where Self: Sized,

Compares and returns the minimum of two values.

fn clamp(self, min: Self, max: Self) -> Self

where Self: Sized,

Restrict a value to a certain interval.

impl PartialEq for OptIn

fn eq(&self, other: &OptIn) -> bool

Equality operator ==.

fn ne(&self, other: &Rhs) -> bool

Inequality operator !=.

impl PartialOrd for OptIn

fn partial_cmp(&self, other: &OptIn) -> Option<Ordering>

This method returns an ordering between self and other values if one exists.

fn lt(&self, other: &Rhs) -> bool

Tests less than (for self and other) and is used by the < operator.

fn le(&self, other: &Rhs) -> bool

Tests less than or equal to (for self and other) and is used by the <= operator.

fn gt(&self, other: &Rhs) -> bool

Tests greater than (for self and other) and is used by the > operator.

fn ge(&self, other: &Rhs) -> bool

Tests greater than or equal to (for self and other) and is used by the >= operator.

impl StructuralPartialEq for OptIn