Struct Neuron

pub struct Neuron<const I: usize> {
    pub weights: [f32; I],
    pub bias: f32,
}

A simple leaky ReLU neuron with I inputs.

The leak slope is 0.01.

Fields

weights: [f32; I]

bias: f32

Implementations

impl<const I: usize> Neuron<I>

pub fn zero() -> Self

Creates a new neuron with weights and bias to 0.

pub fn random_with_0_bias() -> Self

Creates a new neuron with random weights within [-1, 1] and bias to 0.

pub fn random() -> Self

Creates a new neuron with random weights and bias within [-1, 1].

pub fn random_with_range(range: f32) -> Self

Creates a new neuron with random weights and bias within [-range, range].

pub fn output(&self, x: &[f32; I]) -> f32

Computes the neuron’s output for x.

pub fn train<'a>( &mut self, data: impl Iterator<Item = &'a ([f32; I], f32)>, epsilon: f32, )

Trains the neuron from data using back-propagation with epsilon learning rate (per data entry).

Trait Implementations

impl<const I: usize> Clone for Neuron<I>

fn clone(&self) -> Neuron<I>

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl<const I: usize> Debug for Neuron<I>

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

Formats the value using the given formatter.