Enum WeightStrategy 

pub enum WeightStrategy {
    Uniform,
    InverseFrequency {
        class_counts: Option<Vec<usize>>,
    },
    EffectiveNumber {
        beta: f64,
        class_counts: Option<Vec<usize>>,
    },
    Custom(Vec<f32>),
    Curriculum {
        sample_losses: Vec<f32>,
        easy_first: bool,
    },
}

Strategy for computing per-sample weights.

Variants

Uniform

Equal weights for all samples.

InverseFrequency

Inverse frequency weighting (minority classes get higher weight).

Fields

class_counts: Option<Vec<usize>>

Class counts (or will be computed).

EffectiveNumber

Effective number weighting (handles long-tail distributions).

Fields

beta: f64

Beta parameter (typically 0.99 or 0.999).

class_counts: Option<Vec<usize>>

Class counts.

Custom(Vec<f32>)

Custom weights per sample.

Curriculum

Curriculum learning: weight by sample difficulty.

Fields

sample_losses: Vec<f32>

Current loss per sample.

easy_first: bool

Weight easy samples more initially.

Trait Implementations

impl Clone for WeightStrategy

fn clone(&self) -> WeightStrategy

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for WeightStrategy

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

Formats the value using the given formatter.