Enum ParamDistribution 

#[non_exhaustive]
pub enum ParamDistribution {
    Categorical(Vec<ParamValue>),
    Uniform {
        low: f64,
        high: f64,
    },
    LogUniform {
        low: f64,
        high: f64,
    },
    IntUniform {
        low: usize,
        high: usize,
    },
}

A distribution from which hyperparameter values can be sampled.

Used with BayesSearchCV to define a continuous or discrete search space for each hyperparameter.

Examples

use scry_learn::search::ParamDistribution;

let lr = ParamDistribution::LogUniform { low: 0.001, high: 1.0 };
let depth = ParamDistribution::IntUniform { low: 2, high: 10 };

Variants (Non-exhaustive)

Non-exhaustive enums could have additional variants added in future. Therefore, when matching against variants of non-exhaustive enums, an extra wildcard arm must be added to account for any future variants.

Categorical(Vec<ParamValue>)

A set of discrete candidate values (any ParamValue variant).

Uniform

Continuous uniform distribution over [low, high].

Fields

low: f64

Lower bound (inclusive).

high: f64

Upper bound (inclusive).

LogUniform

Log-uniform distribution over [low, high] (sampled in log space). Both low and high must be positive.

Fields

low: f64

Lower bound (inclusive, positive).

high: f64

Upper bound (inclusive, positive).

IntUniform

Discrete uniform distribution over integers [low, high].

Fields

low: usize

Lower bound (inclusive).

high: usize

Upper bound (inclusive).

Trait Implementations

impl Clone for ParamDistribution

fn clone(&self) -> ParamDistribution

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for ParamDistribution

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

Formats the value using the given formatter.