-
Notifications
You must be signed in to change notification settings - Fork 28
/
NeuraNet.pde
172 lines (146 loc) · 4.94 KB
/
NeuraNet.pde
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
class NeuralNet {
int iNodes, hNodes, oNodes, hLayers;
NetMatrix[] weights;
NeuralNet(int input, int hidden, int output, int hiddenLayers) {
iNodes = input;
hNodes = hidden;
oNodes = output;
hLayers = hiddenLayers;
weights = new NetMatrix[hLayers+1];
weights[0] = new NetMatrix(hNodes, iNodes+1);
for(int i=1; i<hLayers; i++) {
weights[i] = new NetMatrix(hNodes,hNodes+1);
}
weights[weights.length-1] = new NetMatrix(oNodes,hNodes+1);
for(NetMatrix w : weights) {
w.randomize();
}
}
void mutate(float mr) {
for(NetMatrix w : weights) {
w.mutate(mr);
}
}
float[] output(float[] inputsArr) {
NetMatrix inputs = weights[0].single_column_net_matrix_from_array(inputsArr);
NetMatrix curr_bias = inputs.add_bias();
for(int i=0; i<hLayers; i++) {
NetMatrix hidden_ip = weights[i].dot(curr_bias);
NetMatrix hidden_op = hidden_ip.activate();
curr_bias = hidden_op.add_bias();
}
NetMatrix output_ip = weights[weights.length-1].dot(curr_bias);
NetMatrix output = output_ip;
return output.to_array();
}
NeuralNet crossover(NeuralNet partner) {
NeuralNet child = new NeuralNet(iNodes,hNodes,oNodes,hLayers);
for(int i=0; i<weights.length; i++) {
child.weights[i] = weights[i].crossover(partner.weights[i]);
}
return child;
}
NeuralNet clone() {
NeuralNet clone = new NeuralNet(iNodes,hNodes,oNodes,hLayers);
for(int i=0; i<weights.length; i++) {
clone.weights[i] = weights[i].clone();
}
return clone;
}
void load(NetMatrix[] weight) {
for(int i=0; i<weights.length; i++) {
weights[i] = weight[i];
}
}
NetMatrix[] pull() {
NetMatrix[] model = weights.clone();
return model;
}
void show(float x, float y, float w, float h, float[] vision, float[] decision) {
float space = 5;
float nSize = (h - (space*(iNodes-2))) / iNodes;
float nSpace = (w - (weights.length*nSize)) / weights.length;
float hBuff = (h - (space*(hNodes-1)) - (nSize*hNodes))/2;
float oBuff = (h - (space*(oNodes-1)) - (nSize*oNodes))/2;
int maxIndex = 0;
for(int i = 1; i < decision.length; i++) {
if(decision[i] > decision[maxIndex]) {
maxIndex = i;
}
}
int lc = 1; //Layer Count
strokeWeight(3);
//DRAW WEIGHTS
for(int i = 0; i < weights[0].rows; i++) { //INPUT TO HIDDEN
for(int j = 0; j < weights[0].cols-1; j++) {
if(weights[0].NetMatrix[i][j] < 0) {
stroke(255,0,0);
} else {
stroke(0,0,255);
}
line(x+nSize,y+(nSize/2)+(j*(space+nSize)),x+nSize+nSpace,y+hBuff+(nSize/2)+(i*(space+nSize)));
}
}
lc++;
for(int a = 1; a < hLayers; a++) {
for(int i = 0; i < weights[a].rows; i++) { //HIDDEN TO HIDDEN
for(int j = 0; j < weights[a].cols-1; j++) {
if(weights[a].NetMatrix[i][j] < 0) {
stroke(255,0,0);
} else {
stroke(0,0,255);
}
line(x+(lc*nSize)+((lc-1)*nSpace),y+hBuff+(nSize/2)+(j*(space+nSize)),x+(lc*nSize)+(lc*nSpace),y+hBuff+(nSize/2)+(i*(space+nSize)));
}
}
lc++;
}
for(int i = 0; i < weights[weights.length-1].rows; i++) { //HIDDEN TO OUTPUT
for(int j = 0; j < weights[weights.length-1].cols-1; j++) {
if(weights[weights.length-1].NetMatrix[i][j] < 0) {
stroke(255,0,0);
} else {
stroke(0,0,255);
}
line(x+(lc*nSize)+((lc-1)*nSpace),y+hBuff+(nSize/2)+(j*(space+nSize)),x+(lc*nSize)+(lc*nSpace),y+oBuff+(nSize/2)+(i*(space+nSize)));
}
}
strokeWeight(1);
lc = 0;
//DRAW NODES
for(int i = 0; i < iNodes; i++) { //DRAW INPUTS
if(vision[i] != 0) {
fill(0,255,0);
} else {
fill(255);
}
stroke(0);
ellipseMode(CORNER);
ellipse(x,y+(i*(nSize+space)),nSize,nSize);
textSize(nSize/2);
textAlign(CENTER,CENTER);
fill(0);
text(i,x+(nSize/2),y+(nSize/2)+(i*(nSize+space)));
}
lc++;
for(int a = 0; a < hLayers; a++) {
for(int i = 0; i < hNodes; i++) { //DRAW HIDDEN
fill(255);
stroke(0);
ellipseMode(CORNER);
ellipse(x+(lc*nSize)+(lc*nSpace),y+hBuff+(i*(nSize+space)),nSize,nSize);
}
lc++;
}
for(int i = 0; i < oNodes; i++) { //DRAW OUTPUTS
fill(255);
stroke(0);
ellipseMode(CORNER);
ellipse(x+(lc*nSpace)+(lc*nSize),y+oBuff+(i*(nSize+space)),nSize,nSize);
}
fill(0);
textSize(15);
textAlign(CENTER,CENTER);
text("Score",x+(lc*nSize)+(lc*nSpace)+nSize/2,y+oBuff+(nSize/2)-2);
}
}