Struct ImportanceStats 

pub struct ImportanceStats {
    pub min: f32,
    pub max: f32,
    pub mean: f32,
    pub std: f32,
    pub sparsity_at_threshold: Vec<(f32, f32)>,
}

Statistical summary of importance scores.

Provides min, max, mean, std for analyzing the distribution of importance values before thresholding.

Fields

min: f32

Minimum importance value

max: f32

Maximum importance value

mean: f32

Mean importance value

std: f32

Standard deviation of importance values

sparsity_at_threshold: Vec<(f32, f32)>

Sparsity achieved at various thresholds: (threshold, sparsity_ratio)

Implementations

impl ImportanceStats

pub fn from_tensor(values: &Tensor) -> Self

Compute statistics from a tensor of importance values.

Arguments

• values - Tensor containing importance scores

Returns

Statistics computed over all values in the tensor.

pub fn sparsity_at(&self, values: &Tensor, threshold: f32) -> f32

Compute sparsity at given threshold.

Sparsity is the fraction of values below the threshold.

Trait Implementations

impl Clone for ImportanceStats

fn clone(&self) -> ImportanceStats

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for ImportanceStats

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

Formats the value using the given formatter.

impl Default for ImportanceStats

fn default() -> Self

Returns the “default value” for a type.