Struct revonet::neuro::NeuralLayer

pub struct NeuralLayer { /* fields omitted */ }

Structure to describe a layer for neural network.

Methods

impl NeuralLayer

fn new(size: usize, actf: ActivationFunctionType) -> NeuralLayer

Create a new layer with given size and activation function.

Arguments:

• size - number of nodes.
• actf - type of activation function.

fn init_weights<R: Rng + Sized>(&mut self, inputs_num: usize, rng: &mut R)

Initializes weights of the layer.

Arguments:

• inputs_num - number of inputs for the layer.
• rng - mutable reference to the external RNG.

fn compute(&mut self, xs: &[f32])

Compute output of the layer given a vector of input signals. To get outputs use get_outputs function.

Arguments:

• xs -- input vector.

fn get_inputs_num(&self) -> usize

Returns number of inputs.

fn len(&self) -> usize

Return number of nodes in the layer.

fn get_outputs(&self) -> &[f32]

Returns references to the slice, containing values of current outputs.

fn get_weights(&self) -> (Vec<f32>, Vec<f32>)

Return flattened vector of weights and biases.

fn set_weights(&mut self, ws: &Vec<f32>, bs: &Vec<f32>)

Set weights and biases for the layer.

Arguments:

• ws - vector of flattened weights.
• bs - vector of biases.

Trait Implementations

impl Clone for NeuralLayer

fn clone(&self) -> NeuralLayer

Returns a copy of the value.

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

Performs copy-assignment from source.

impl Debug for NeuralLayer

fn fmt(&self, __arg_0: &mut Formatter) -> Result

Formats the value using the given formatter.