mic::neural_nets::unit_tests::Conv5x6x1Filter1x4x4s1Float Class Reference

Test Fixture - layer of input size 5x6x1 and with filter bank of 1 filter of size 4x4 with stride 1, floats. Math example taken from: http://soumith.ch/ex/pages/2014/08/07/why-rotate-weights-convolution-gradient/. More...

#include <Convolution_tests.hpp>

Inheritance diagram for mic::neural_nets::unit_tests::Conv5x6x1Filter1x4x4s1Float:

Collaboration diagram for mic::neural_nets::unit_tests::Conv5x6x1Filter1x4x4s1Float:

Public Member Functions
	Conv5x6x1Filter1x4x4s1Float ()

Protected Member Functions
virtual void	SetUp ()

Private Attributes
mic::mlnn::convolution::Convolution < float >	layer
	Object to be tested. More...
mic::types::MatrixPtr< float >	x
	Test x - used in forward pass. More...
mic::types::MatrixPtr< float >	desired_y
	Desired output for a given x. More...
mic::types::MatrixPtr< float >	dy
	Gradient passed to backpropagation. More...
mic::types::MatrixPtr< float >	desired_dx
	Desired gradient dy from backpropagation. More...

Detailed Description

Test Fixture - layer of input size 5x6x1 and with filter bank of 1 filter of size 4x4 with stride 1, floats. Math example taken from: http://soumith.ch/ex/pages/2014/08/07/why-rotate-weights-convolution-gradient/.

Author

tkornuta

Definition at line 506 of file Convolution_tests.hpp.

Constructor & Destructor Documentation

mic::neural_nets::unit_tests::Conv5x6x1Filter1x4x4s1Float::Conv5x6x1Filter1x4x4s1Float ( )

inline

Definition at line 509 of file Convolution_tests.hpp.

References desired_dx, desired_y, dy, and x.

Member Function Documentation

virtual void mic::neural_nets::unit_tests::Conv5x6x1Filter1x4x4s1Float::SetUp ( )

inlineprotectedvirtual

Definition at line 521 of file Convolution_tests.hpp.

References layer, mic::mlnn::Layer< eT >::p, and x.

Member Data Documentation

mic::types::MatrixPtr<float> mic::neural_nets::unit_tests::Conv5x6x1Filter1x4x4s1Float::desired_dx

private

Desired gradient dy from backpropagation.

Definition at line 560 of file Convolution_tests.hpp.

Referenced by Conv5x6x1Filter1x4x4s1Float().

mic::types::MatrixPtr<float> mic::neural_nets::unit_tests::Conv5x6x1Filter1x4x4s1Float::desired_y

private

Desired output for a given x.

Definition at line 554 of file Convolution_tests.hpp.

Referenced by Conv5x6x1Filter1x4x4s1Float().

mic::types::MatrixPtr<float> mic::neural_nets::unit_tests::Conv5x6x1Filter1x4x4s1Float::dy

private

Gradient passed to backpropagation.

Definition at line 557 of file Convolution_tests.hpp.

Referenced by Conv5x6x1Filter1x4x4s1Float().

mic::mlnn::convolution::Convolution<float> mic::neural_nets::unit_tests::Conv5x6x1Filter1x4x4s1Float::layer

private

Object to be tested.

Definition at line 548 of file Convolution_tests.hpp.

Referenced by SetUp().

mic::types::MatrixPtr<float> mic::neural_nets::unit_tests::Conv5x6x1Filter1x4x4s1Float::x

private

Test x - used in forward pass.

Definition at line 551 of file Convolution_tests.hpp.

Referenced by Conv5x6x1Filter1x4x4s1Float(), and SetUp().

---

The documentation for this class was generated from the following file:

• src/mlnn/convolution/Convolution_tests.hpp