deepboof.impl.backward.standard

Class NumericalGradient_F64

java.lang.Object

deepboof.impl.backward.standard.NumericalGradient_F64

All Implemented Interfaces:

NumericalGradient<deepboof.tensors.Tensor_F64>

public class NumericalGradient_F64
extends java.lang.Object
implements NumericalGradient<deepboof.tensors.Tensor_F64>

Implementation of NumericalGradient for Tensor_F64

Constructor Summary

Constructors

Constructor and Description
NumericalGradient_F64()

Method Summary

Modifier and Type	Method and Description
void	configure(double T) Overrides default settings for computing numerical gradient.
void	differentiate(deepboof.tensors.Tensor_F64 input, java.util.List<deepboof.tensors.Tensor_F64> parameters, deepboof.tensors.Tensor_F64 dout, deepboof.tensors.Tensor_F64 gradientInput, java.util.List<deepboof.tensors.Tensor_F64> gradientParameters) Performs numerical differentiation to compute the gradients of input and parameters.
void	setFunction(deepboof.Function<deepboof.tensors.Tensor_F64> function) Sets the function which will be differentiated and other parameters.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

NumericalGradient_F64

public NumericalGradient_F64()

Method Detail

configure

public void configure(double T)

Description copied from interface: NumericalGradient

Overrides default settings for computing numerical gradient.

Specified by:

configure in interface NumericalGradient<deepboof.tensors.Tensor_F64>

Parameters:

T - Sampling distance used for numerical differentiation

setFunction

public void setFunction(deepboof.Function<deepboof.tensors.Tensor_F64> function)

Description copied from interface: NumericalGradient

Sets the function which will be differentiated and other parameters. Function.initialize(int...) should have already been called.

Specified by:

setFunction in interface NumericalGradient<deepboof.tensors.Tensor_F64>

Parameters:

function - The function which is to be differentiated

differentiate

public void differentiate(deepboof.tensors.Tensor_F64 input,
                          java.util.List<deepboof.tensors.Tensor_F64> parameters,
                          deepboof.tensors.Tensor_F64 dout,
                          deepboof.tensors.Tensor_F64 gradientInput,
                          java.util.List<deepboof.tensors.Tensor_F64> gradientParameters)

Description copied from interface: NumericalGradient

Performs numerical differentiation to compute the gradients of input and parameters. When numerical differentiation is being performed input and parameters will be modified and then returned to their original state.

Specified by:

differentiate in interface NumericalGradient<deepboof.tensors.Tensor_F64>

Parameters:

input - The same input tensor which was passed in during the forward pass.

parameters - The same parameters which was passed in during the forward pass.

dout - Derivative of output, computed from next layer.

gradientInput - Storage for gradient of input

gradientParameters - Storage for gradients of parameters