Struct SimpleNeuralNetwork 

pub struct SimpleNeuralNetwork<const I: usize, const H: usize, const O: usize> { /* private fields */ }

A simple feedforward neural network with I inputs, H neurons in the hidden layer, and O outputs.

Implementations

impl<const I: usize, const H: usize, const O: usize> SimpleNeuralNetwork<I, H, O>

pub fn new(learning_rate: f32) -> Self

Creates a new SimpleNeuralNetwork with random weights and biases.

Trait Implementations

impl<const I: usize, const H: usize, const O: usize> NeuralNetwork<I, O> for SimpleNeuralNetwork<I, H, O>

type LayerOutputs = SimpleLayerOutputs<H, O>

The type representing the outputs of each layer in the network.

fn feedforward(&self, input: &RowVector<I>) -> Self::LayerOutputs

Feedforward the input through the network, updating layer outputs, and returning the final output.

fn backpropagate( &mut self, input: &RowVector<I>, output: &RowVector<O>, layer_outputs: Self::LayerOutputs, ) -> f32

Perform backpropagation to adjust weights and biases based on the target output, returning the loss.

fn train_once<BI, BO>(&mut self, input: BI, output: BO) -> f32

where BI: Borrow<RowVector<I>>, BO: Borrow<RowVector<O>>,

Train the neural network once with a pair of input and output, returning the average loss.

fn train<D, T>( &mut self, data: D, callback: Option<impl FnMut(usize, f32)>, ) -> (usize, f32)

where D: IntoIterator<Item = T>, T: Borrow<(RowVector<I>, RowVector<O>)>,

Train the neural network with the provided data, calling an optional callback after each sample, returning the total number of samples and the final average loss.

fn predict(&self, input: &RowVector<I>) -> RowVector<O>

Predict the output for a given input.