Trait Backward 

pub trait Backward<X, Delta = X> {
    type Elem;

    // Required method
    fn backward(&mut self, input: &X, delta: &Delta, gamma: Self::Elem);
}

The Backward trait establishes a common interface for completing a single backward step in a neural network or machine learning model.

Required Associated Types

type Elem

Required Methods

fn backward(&mut self, input: &X, delta: &Delta, gamma: Self::Elem)

Implementations on Foreign Types

impl<A, S, D, S1, D1, S2, D2> Backward<ArrayBase<S1, D1, A>, ArrayBase<S2, D2, A>> for ArrayBase<S, D, A>

where A: 'static + Copy + Num, D: Dimension, S: DataMut<Elem = A>, D1: Dimension, D2: Dimension, S1: Data<Elem = A>, S2: Data<Elem = A>, for<'b> &'b ArrayBase<S1, D1, A>: Dot<ArrayView<'b, A, D2>, Output = Array<A, D2>>,

type Elem = A

fn backward( &mut self, input: &ArrayBase<S1, D1, A>, delta: &ArrayBase<S2, D2, A>, gamma: Self::Elem, )

Implementors