Trait goko::plugins::gaussians::ContinousDistribution

pub trait ContinousDistribution: Clone + 'static {
    fn ln_pdf<T: PointRef>(&self, point: &T) -> Option<f64>;
    fn sample<R: Rng>(&self, rng: &mut R) -> Vec<f32>;
    fn kl_divergence(&self, other: &Self) -> Option<f64>;
}

Required methods

fn ln_pdf<T: PointRef>(&self, point: &T) -> Option<f64>

Pass none if you want to test for a singleton, returns 0 if

fn sample<R: Rng>(&self, rng: &mut R) -> Vec<f32>

Samples a point from this distribution

fn kl_divergence(&self, other: &Self) -> Option<f64>

Computes the KL divergence of two bucket probs. KL(self || other) Returns None if the support of the self is not a subset of the support of the other, or the calculation is undefined.

Loading content...

Implementors

impl ContinousDistribution for DiagGaussian

fn ln_pdf<T: PointRef>(&self, point: &T) -> Option<f64>

fn sample<R: Rng>(&self, rng: &mut R) -> Vec<f32>

fn kl_divergence(&self, other: &DiagGaussian) -> Option<f64>

Loading content...