Struct PrototypicalNetwork 

pub struct PrototypicalNetwork { /* private fields */ }

Prototypical Networks for few-shot learning (Snell et al., 2017).

Learns a metric space where classification is performed by computing distances to class prototypes (mean embeddings of support examples).

Implementations

impl PrototypicalNetwork

pub fn new(distance: DistanceMetric) -> PrototypicalNetwork

pub fn compute_prototypes( &self, support: &[(Vec<f32>, usize)], ) -> Vec<(usize, Vec<f32>)>

Compute class prototypes from support set embeddings. support: Vec of (embedding, class_label)

pub fn classify(&self, query: &[f32], prototypes: &[(usize, Vec<f32>)]) -> usize

Classify query embedding against prototypes.

pub fn predict_proba( &self, query: &[f32], prototypes: &[(usize, Vec<f32>)], ) -> Vec<(usize, f32)>

Compute class probabilities (softmax of negative distances).

Trait Implementations

impl Clone for PrototypicalNetwork

fn clone(&self) -> PrototypicalNetwork

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for PrototypicalNetwork

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

Formats the value using the given formatter.

impl Default for PrototypicalNetwork

fn default() -> PrototypicalNetwork

Returns the “default value” for a type.