Struct TrainableKernelMixture 

pub struct TrainableKernelMixture { /* private fields */ }

Trainable adapter around a LearnedMixtureKernel.

Implementations

impl TrainableKernelMixture

pub fn new(inner: LearnedMixtureKernel) -> Self

Wrap an existing mixture kernel.

pub fn num_parameters(&self) -> usize

Number of trainable logits.

pub fn parameters(&self) -> &[f64]

Borrow the trainable parameters (logits).

pub fn weights(&self) -> Vec<f64>

Mixture weights after softmax.

pub fn evaluate(&self, x: &[f64], y: &[f64]) -> Result<f64>

Forward pass — mixture kernel value.

pub fn evaluate_with_gradient( &self, x: &[f64], y: &[f64], ) -> Result<(f64, Vec<f64>)>

Forward + gradient in a single pass.

pub fn gradient(&self, x: &[f64], y: &[f64]) -> Result<Vec<f64>>

Pure gradient (no forward re-use).

pub fn step(&mut self, gradient: &[f64], learning_rate: f64) -> Result<()>

Apply a vanilla gradient-descent step w <- w - lr * g.

pub fn set_parameters(&mut self, new_logits: Vec<f64>) -> Result<()>

Replace the logits outright (useful for optimizer-driven updates).

pub fn inner(&self) -> &LearnedMixtureKernel

Borrow the underlying mixture kernel (read-only).

pub fn into_inner(self) -> LearnedMixtureKernel

Consume the adapter and return the underlying mixture kernel.

Trait Implementations

impl Clone for TrainableKernelMixture

fn clone(&self) -> TrainableKernelMixture

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for TrainableKernelMixture

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

Formats the value using the given formatter.

impl From<LearnedMixtureKernel> for TrainableKernelMixture

fn from(inner: LearnedMixtureKernel) -> Self

Converts to this type from the input type.