Struct MatchingNetwork 

pub struct MatchingNetwork { /* private fields */ }

Matching network for few-shot learning.

Uses attention mechanism to match query examples to support examples.

Implementations

impl MatchingNetwork

pub fn new(metric: DistanceMetric) -> Self

Create a new matching network.

pub fn set_support(&mut self, support: SupportSet)

Set support set.

pub fn compute_attention( &self, query: &ArrayView1<'_, f64>, ) -> TrainResult<Array1<f64>>

Compute attention weights between query and all support examples.

Arguments

• query - Query feature vector

Returns

Attention weights for each support example

pub fn predict_proba( &self, query: &ArrayView1<'_, f64>, ) -> TrainResult<Array1<f64>>

Predict class using attention-weighted voting.

Arguments

• query - Query feature vector

Returns

Predicted class probabilities

pub fn predict(&self, query: &ArrayView1<'_, f64>) -> TrainResult<usize>

Predict class label.

Trait Implementations

impl Clone for MatchingNetwork

fn clone(&self) -> MatchingNetwork

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for MatchingNetwork

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

Formats the value using the given formatter.