UESMANN CPP  1.0
Reference implementation of UESMANN
Functions
Core functionality tests
Collaboration diagram for Core functionality tests:

Functions

 BOOST_AUTO_TEST_CASE (example)
 Test the basic example. More...
 
 BOOST_AUTO_TEST_CASE (subset)
 Test that subsetting examples works. More...
 
template<class T >
void sshuffle (T *x, int ct)
 simple shuffle for testing - performs a Fisher-Yates shuffle on an array of items of class T. More...
 
 BOOST_AUTO_TEST_CASE (alt)
 Test the alternate() function. More...
 
 BOOST_AUTO_TEST_CASE (altex)
 Test the alternation function on examples, simple version. More...
 
 BOOST_AUTO_TEST_CASE (stride)
 test strided example shuffle, 4 different modulator levels More...
 
 BOOST_AUTO_TEST_CASE (altex4)
 test strided example shuffle, 4 different modulator levels More...
 
void zero (Net *n)
 set all parameters (weights and biases) in a network to zero More...
 
 BOOST_AUTO_TEST_CASE (testmse)
 Test mean sum squared error of outputs. This test finds the mean of the sum of the squared errors on the outputs across all examples in the test set, in the case of a network where all the parameters are zero. In this case, all outputs will be 0.5 given the logistic sigmoid activation function. We test for the correct value determined by lots of print statements during development. More...
 
 BOOST_AUTO_TEST_CASE (loadmnist)
 Loading MNIST data and converting to an example set. Ensure we can load MNIST data into an example set, and that the image and its label are correct. The former is hard to test automatically, so I'll rely on having eyeballed it. More...
 

Detailed Description

Function Documentation

BOOST_AUTO_TEST_CASE ( example  )

Test the basic example.

Definition at line 57 of file testBasic.cpp.

BOOST_AUTO_TEST_CASE ( subset  )

Test that subsetting examples works.

Definition at line 76 of file testBasic.cpp.

BOOST_AUTO_TEST_CASE ( alt  )

Test the alternate() function.

Definition at line 120 of file testBasic.cpp.

BOOST_AUTO_TEST_CASE ( altex  )

Test the alternation function on examples, simple version.

Definition at line 154 of file testBasic.cpp.

BOOST_AUTO_TEST_CASE ( stride  )

test strided example shuffle, 4 different modulator levels

Definition at line 172 of file testBasic.cpp.

BOOST_AUTO_TEST_CASE ( altex4  )

test strided example shuffle, 4 different modulator levels

Definition at line 222 of file testBasic.cpp.

BOOST_AUTO_TEST_CASE ( testmse  )

Test mean sum squared error of outputs. This test finds the mean of the sum of the squared errors on the outputs across all examples in the test set, in the case of a network where all the parameters are zero. In this case, all outputs will be 0.5 given the logistic sigmoid activation function. We test for the correct value determined by lots of print statements during development.

Definition at line 299 of file testBasic.cpp.

BOOST_AUTO_TEST_CASE ( loadmnist  )

Loading MNIST data and converting to an example set. Ensure we can load MNIST data into an example set, and that the image and its label are correct. The former is hard to test automatically, so I'll rely on having eyeballed it.

Definition at line 322 of file testBasic.cpp.

template<class T >
void sshuffle ( T *  x,
int  ct 
)

simple shuffle for testing - performs a Fisher-Yates shuffle on an array of items of class T.

Definition at line 104 of file testBasic.cpp.

void zero ( Net n)

set all parameters (weights and biases) in a network to zero

Parameters
nthe network to zero

Definition at line 282 of file testBasic.cpp.