forward

Documentation

Index

• Variables
• type Layer
• type Network
  • func (net Network) Activate(inputs evo.Matrix) (outputs evo.Matrix, err error)
• type Translator
  • func (t Translator) Translate(sub evo.Substrate) (net evo.Network, err error)

Variables

var (
	ErrNoSensors = errors.New("network requires at least 1 input neuron")
	ErrNoOutputs = errors.New("network requires at least 1 output neuron")
)

Known errors

Types

type Layer

type Layer struct {
	Activations []evo.Activation // activation functions for neurons in this layer
	Biases      []float64        // bias values for this layer
	Sources     []int            // indexes of source layers
	Weights     []*mat.Dense     // weights to each source layer
}

Layer in the neural network

type Network

type Network struct {
	Layers []Layer
}

Network of layers

func (Network) Activate

func (net Network) Activate(inputs evo.Matrix) (outputs evo.Matrix, err error)

Activate activates the network using the incoming matrix of values

type Translator

type Translator struct {
	DisableSortCheck bool
}

Translator transforms substrates into networks

func (Translator) Translate

func (t Translator) Translate(sub evo.Substrate) (net evo.Network, err error)

Translate the substrate into a network