nn

README

Neural Network Package

WIP

Documentation

Index

• func Apply(fn func(x float64) float64, input *mat.Dense) *mat.Dense
• func Arctan(input *mat.Dense) *mat.Dense
• func ArctanDeriv(input *mat.Dense) *mat.Dense
• func Gaussian(input *mat.Dense) *mat.Dense
• func GaussianDeriv(input *mat.Dense) *mat.Dense
• func Hmse(y_true, y_pred *mat.Dense, params []float64) float64
• func HmseDerivative(y_true, y_pred *mat.Dense, params []float64) *mat.Dense
• func Huber(y_true, y_pred *mat.Dense, params []float64) float64
• func HuberDerivative(y_true, y_pred *mat.Dense, params []float64) *mat.Dense
• func Linear(input *mat.Dense) *mat.Dense
• func LinearDeriv(input *mat.Dense) *mat.Dense
• func Mae(y_true, y_pred *mat.Dense, params []float64) float64
• func MaeDerivative(y_true, y_pred *mat.Dense, params []float64) *mat.Dense
• func Mse(y_true, y_pred *mat.Dense, params []float64) float64
• func MseDerivative(y_true, y_pred *mat.Dense, params []float64) *mat.Dense
• func ReLu(input *mat.Dense) *mat.Dense
• func ReLuDeriv(input *mat.Dense) *mat.Dense
• func Rmse(y_true, y_pred *mat.Dense, params []float64) float64
• func RmseDerivative(y_true, y_pred *mat.Dense, params []float64) *mat.Dense
• func Sigmoid(input *mat.Dense) *mat.Dense
• func SigmoidDeriv(input *mat.Dense) *mat.Dense
• func Tanh(input *mat.Dense) *mat.Dense
• func TanhDeriv(input *mat.Dense) *mat.Dense
• type Activation
• type ActivationLayer
  • func InitActivationLayer(activation Activation, activationDeriv Activation) *ActivationLayer
  • func (layer *ActivationLayer) BackPropagation(output_error *mat.Dense, learning_rate float64) *mat.Dense
  • func (layer *ActivationLayer) ForwardPropagation(input *mat.Dense) *mat.Dense
• type ConvolutionalLayer
  • func InitConvolutionalLayer(input_size int, output_size int) *ConvolutionalLayer
  • func (layer *ConvolutionalLayer) BackPropagation(output_error *mat.Dense, learning_rate float64) *mat.Dense
  • func (layer *ConvolutionalLayer) ForwardPropagation(input *mat.Dense) *mat.Dense
• type DenseLayer
  • func InitDenseLayer(input_size int, output_size int) *DenseLayer
  • func (layer *DenseLayer) BackPropagation(output_error *mat.Dense, learning_rate float64) *mat.Dense
  • func (layer *DenseLayer) ForwardPropagation(input *mat.Dense) *mat.Dense
• type FlattenLayer
  • func InitFlattenLayer(input_size int, output_size int) *FlattenLayer
  • func (layer *FlattenLayer) BackPropagation(output_error *mat.Dense, learning_rate float64) *mat.Dense
  • func (layer *FlattenLayer) ForwardPropagation(input *mat.Dense) *mat.Dense
  • func (layer *FlattenLayer) GetShape() (int, int)
• type Layer
• type Network
  • func (net *Network) AddLayer(layerType, activationType string, output_nodes int)
  • func (net *Network) Fit(x_train, y_train [][]float64, epochs int, learning_rate float64, debug bool)
  • func (net *Network) Predict(input [][]float64) [][]float64
  • func (net *Network) SetLoss(lossType string, params []float64)

Functions

func Apply

func Apply(fn func(x float64) float64, input *mat.Dense) *mat.Dense

*

• Apply()
• Helper that takes a function as a parameter to perform on the layer input

func Arctan

func Arctan(input *mat.Dense) *mat.Dense

*

• Arctan Activation Funcs

func ArctanDeriv

func ArctanDeriv(input *mat.Dense) *mat.Dense

func Gaussian

func Gaussian(input *mat.Dense) *mat.Dense

*

• Gaussian Activation Funcs

func GaussianDeriv

func GaussianDeriv(input *mat.Dense) *mat.Dense

func Hmse

func Hmse(y_true, y_pred *mat.Dense, params []float64) float64

*

• Half Mean Squared Error Loss Functions

func HmseDerivative

func HmseDerivative(y_true, y_pred *mat.Dense, params []float64) *mat.Dense

func Huber

func Huber(y_true, y_pred *mat.Dense, params []float64) float64

*

• Huber Loss Functions
• Implements a pseudo-huber loss type to approximate
• Takes in a delta value as a parameter

func HuberDerivative

func HuberDerivative(y_true, y_pred *mat.Dense, params []float64) *mat.Dense

func Linear

func Linear(input *mat.Dense) *mat.Dense

*

• Linear Activation Funcs

func LinearDeriv

func LinearDeriv(input *mat.Dense) *mat.Dense

func Mae

func Mae(y_true, y_pred *mat.Dense, params []float64) float64

*

• Mean Absolute Error Loss Functions
• Derivative defaults to 1 when y_pred = y_true due to not being differentiable

func MaeDerivative

func MaeDerivative(y_true, y_pred *mat.Dense, params []float64) *mat.Dense

func Mse

func Mse(y_true, y_pred *mat.Dense, params []float64) float64

*

• Mean Squared Error Loss Functions

func MseDerivative

func MseDerivative(y_true, y_pred *mat.Dense, params []float64) *mat.Dense

func ReLu

func ReLu(input *mat.Dense) *mat.Dense

*

• ReLu Activation Funcs
• Derivative defaults to 0 if x = 0

func ReLuDeriv

func ReLuDeriv(input *mat.Dense) *mat.Dense

func Rmse

func Rmse(y_true, y_pred *mat.Dense, params []float64) float64

*

• Root Mean Squared Error Loss Functions

func RmseDerivative

func RmseDerivative(y_true, y_pred *mat.Dense, params []float64) *mat.Dense

func Sigmoid

func Sigmoid(input *mat.Dense) *mat.Dense

*

• Sigmoid Activation Funcs

func SigmoidDeriv

func SigmoidDeriv(input *mat.Dense) *mat.Dense

func Tanh

func Tanh(input *mat.Dense) *mat.Dense

*

• Tanh Activation Funcs

func TanhDeriv

func TanhDeriv(input *mat.Dense) *mat.Dense

Types

type Activation

type Activation func(input *mat.Dense) *mat.Dense

*

• Activation Function
• Defines format of any activation function

type ActivationLayer

type ActivationLayer struct {
	INPUT           *mat.Dense
	OUTPUT          *mat.Dense
	ACTIVATION      Activation
	ACTIVATIONDERIV Activation
}

*

• ActivationLayer Struct
• Contains the activation function and its derivative for the layer

func InitActivationLayer

func InitActivationLayer(activation Activation, activationDeriv Activation) *ActivationLayer

*

• InitActivationLayer()
• Initializes an activation the activation function and its derivative

func (*ActivationLayer) BackPropagation

func (layer *ActivationLayer) BackPropagation(output_error *mat.Dense, learning_rate float64) *mat.Dense

*

• BackPropagation()
• Performs back propagation for an activation layer required by Layer interface
• Returns matrix of error generated from derivative of activation function

func (*ActivationLayer) ForwardPropagation

func (layer *ActivationLayer) ForwardPropagation(input *mat.Dense) *mat.Dense

*

• ForwardPropagation()
• Performs forward propagation for an activation layer required by Layer interface
• Returns output matrix with the function performed

type ConvolutionalLayer

type ConvolutionalLayer struct {
	INPUT  *mat.Dense
	OUTPUT *mat.Dense
}

*

• ConvolutionalLayer Struct
• WIP

func InitConvolutionalLayer

func InitConvolutionalLayer(input_size int, output_size int) *ConvolutionalLayer

*

• InitConvolutionalLayer()
• WIP

func (*ConvolutionalLayer) BackPropagation

func (layer *ConvolutionalLayer) BackPropagation(output_error *mat.Dense, learning_rate float64) *mat.Dense

*

• BackPropagation()
• Performs back propagation for a convolutional layer required by Layer interface
• WIP

func (*ConvolutionalLayer) ForwardPropagation

func (layer *ConvolutionalLayer) ForwardPropagation(input *mat.Dense) *mat.Dense

*

• ForwardPropagation()
• Performs forward propagation for a convolutional layer required by Layer interface
• WIP

type DenseLayer

type DenseLayer struct {
	INPUT   *mat.Dense
	OUTPUT  *mat.Dense
	WEIGHTS *mat.Dense
	BIAS    *mat.Dense
}

*

• DenseLayer Struct
• Contains weights and bias for a dense / fully connected layer in a neural network

func InitDenseLayer

func InitDenseLayer(input_size int, output_size int) *DenseLayer

*

• InitDenseLayer()
• Initializes a dense layer given input and output shape

func (*DenseLayer) BackPropagation

func (layer *DenseLayer) BackPropagation(output_error *mat.Dense, learning_rate float64) *mat.Dense

*

• BackPropagation()
• Performs back propagation for a dense layer required by Layer interface
• Returns matrix of the input error from the current layer

func (*DenseLayer) ForwardPropagation

func (layer *DenseLayer) ForwardPropagation(input *mat.Dense) *mat.Dense

*

• ForwardPropagation()
• Performs forward propagation for a dense layer required by Layer interface
• Returns dot product of input and weights plus the bias

type FlattenLayer

type FlattenLayer struct {
	INPUT        *mat.Dense
	OUTPUT       *mat.Dense
	INPUT_SHAPE  int
	OUTPUT_SHAPE int
}

*

• FlattenLayer Struct
• Contains input and output matrices

func InitFlattenLayer

func InitFlattenLayer(input_size int, output_size int) *FlattenLayer

*

• InitFlattenLayer()
• Flattened layer serves to simply reshape inputs and outputs

func (*FlattenLayer) BackPropagation

func (layer *FlattenLayer) BackPropagation(output_error *mat.Dense, learning_rate float64) *mat.Dense

*

• BackPropagation()
• Performs back propagation for a flat layer required by Layer interface
• WIP for Convolutional Layer

func (*FlattenLayer) ForwardPropagation

func (layer *FlattenLayer) ForwardPropagation(input *mat.Dense) *mat.Dense

*

• ForwardPropagation()
• Performs forward propagation for a flat layer required by Layer interface
• WIP for Convolutional Layer

func (*FlattenLayer) GetShape

func (layer *FlattenLayer) GetShape() (int, int)

*

• GetShape()
• Accessor for flat layer shape

type Layer

type Layer interface {
	ForwardPropagation(input *mat.Dense) *mat.Dense
	BackPropagation(output_error *mat.Dense, learning_rate float64) *mat.Dense
}

*

• Layer Interface
• Defines required functions for any layer

type Network

type Network struct {
	LAYERS      []*Layer
	LOSSFUNC    func(y_true, y_pred *mat.Dense, params []float64) float64
	LOSSDERIV   func(y_true, y_pred *mat.Dense, params []float64) *mat.Dense
	LOSSPARAMS  []float64
	OUTPUT_SIZE int
	LAYER_COUNT int
}

*

• Network Struct
• Contains neural network specifics and helper functions
• Defines layers and loss functions

func (*Network) AddLayer

func (net *Network) AddLayer(layerType, activationType string, output_nodes int)

*

• AddLayer()
• Adds specific layer to neural network given activation type and input/output neuron dimensions

func (*Network) Fit

func (net *Network) Fit(x_train, y_train [][]float64, epochs int, learning_rate float64, debug bool)

*

• Fit()
• Fit the neural network given training data, number of epochs, and a learning rate
• Debug option provided for printing data from each epoch

func (*Network) Predict

func (net *Network) Predict(input [][]float64) [][]float64

*

• Predict()
• Given sample data, return predicted values ran through the neural network

func (*Network) SetLoss

func (net *Network) SetLoss(lossType string, params []float64)

*

• SetLoss()
• Contains neural network specifics and helper functions
• Defines layers and loss functions