forked from einsnull/DeepLearningToolBox
-
Notifications
You must be signed in to change notification settings - Fork 0
Expand file tree
/
Copy pathFunctionBase.h
More file actions
179 lines (161 loc) · 3.43 KB
/
FunctionBase.h
File metadata and controls
179 lines (161 loc) · 3.43 KB
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
#pragma once
#include <Eigen/Dense>
using namespace Eigen;
using namespace std;
//scalar reciprocal
double reciprocalScalar(double x)
{
return 1.0/x;
}
//scalar sigmoid function
double sigmoidScalar(double x)
{
return 1.0 / (1 + exp(-x));
}
//scalar log function
double logScalar(double x)
{
return log(x);
}
//scalar exp function
double expScalar(double x)
{
return exp(x);
}
//scalar sqrt function
double sqrtScalar(double x)
{
return sqrt(x);
}
class FunctionBase
{
public:
MatrixXd bsxfunMinus(MatrixXd &m,MatrixXd &x);
MatrixXd bsxfunRDivide(MatrixXd &m,MatrixXd &x);
MatrixXd bsxfunPlus(MatrixXd &m,MatrixXd &x);
MatrixXd sigmoid(MatrixXd &z);
MatrixXd sigmoidGradient(MatrixXd &z);
MatrixXd binaryCols(MatrixXi &labels,int numOfClasses);
MatrixXd expMat(MatrixXd &z);
MatrixXd logMat(MatrixXd &z);
MatrixXd sqrtMat(MatrixXd &z);
MatrixXd reciprocal(MatrixXd &z);
double calcAccurancy(MatrixXi &pred,MatrixXi &labels);
FunctionBase();
};
FunctionBase::FunctionBase()
{
#ifdef _WINDOWS_
//set eigen threads
SYSTEM_INFO info;
GetSystemInfo(&info);
Eigen::setNbThreads(info.dwNumberOfProcessors);
#endif
}
//calculate m minus broadcasted x
MatrixXd FunctionBase::bsxfunMinus(MatrixXd &m,MatrixXd &x)
{
MatrixXd r = m;
if(x.rows() == 1)
{
r = x.replicate(m.rows(),1);
}
if(x.cols() == 1)
{
r = x.replicate(1,m.cols());
}
return m - r;
}
//calculate m right divide broadcasted x
MatrixXd FunctionBase::bsxfunRDivide(MatrixXd &m,MatrixXd &x)
{
MatrixXd r = m;
if(x.rows() == 1)
{
r = x.replicate(m.rows(),1);
}
if(x.cols() == 1)
{
r = x.replicate(1,m.cols());
}
return m.cwiseQuotient(r);
}
//calculate m plus broadcasted x
MatrixXd FunctionBase::bsxfunPlus(MatrixXd &m,MatrixXd &x)
{
MatrixXd r = m;
if(x.rows() == 1)
{
r = x.replicate(m.rows(),1);
}
if(x.cols() == 1)
{
r = x.replicate(1,m.cols());
}
return m + r;
}
//The gradient of sigmoid function
MatrixXd FunctionBase::sigmoidGradient(MatrixXd &z)
{
//return sigmoid(z) .* (1 - sigmoid(z))
MatrixXd result;
MatrixXd sigm = sigmoid(z);
MatrixXd item = MatrixXd::Ones(z.rows(),z.cols()) - sigm;
result = sigm.cwiseProduct(item);
return result;
}
//component wise sigmoid function
MatrixXd FunctionBase::sigmoid(MatrixXd &z)
{
return z.unaryExpr(ptr_fun(sigmoidScalar));
}
//component wise sqrt function
MatrixXd FunctionBase::sqrtMat(MatrixXd &z)
{
return z.unaryExpr(ptr_fun(sqrtScalar));
}
//return binary code of labels
MatrixXd FunctionBase::binaryCols(MatrixXi &labels,int numOfClasses)
{
// return binary code of labels
//eye function
MatrixXd e = MatrixXd::Identity(numOfClasses,numOfClasses);
int numOfExamples = labels.rows();
int inputSize = e.cols();
MatrixXd result(inputSize,numOfExamples);
for(int i = 0; i < numOfExamples; i++)
{
int idx = labels(i,0);
result.col(i) = e.col(idx);
}
return result;
}
//component wise exp function
MatrixXd FunctionBase::expMat(MatrixXd &z)
{
return z.unaryExpr(ptr_fun(expScalar));
}
//component wise log function
MatrixXd FunctionBase::logMat(MatrixXd &z)
{
return z.unaryExpr(ptr_fun(logScalar));
}
//calculate the similarity of two matrix
double FunctionBase::calcAccurancy(MatrixXi &pred,MatrixXi &labels)
{
int numOfExamples = pred.rows();
double sum = 0;
for(int i = 0; i < numOfExamples; i++)
{
if(pred(i,0) == labels(i,0))
{
sum += 1;
}
}
return sum / numOfExamples;
}
//return 1.0 ./ z
MatrixXd FunctionBase::reciprocal(MatrixXd &z)
{
return z.unaryExpr(ptr_fun(reciprocalScalar));
}