/
NN.py
executable file
·106 lines (93 loc) · 3.26 KB
/
NN.py
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
import torch
import torch.nn as nn
# Device configuration
device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
# seed for weight initialization
torch.manual_seed(1234)
# Fully connected neural network
class NeuralNet(nn.Module):
def __init__(self, input_size, units, output_size):
super(NeuralNet, self).__init__()
self.input = nn.Linear(input_size, units[0])
self.relu = nn.ReLU()
self.fcs = []
if(len(units) > 1):
for i in range(0, len(units)-1):
self.fcs.append(nn.Linear(units[i], units[i+1]))
self.output = nn.Linear(units[-1], output_size)
def forward(self, x):
out = self.input(x)
out = self.relu(out)
if(len(self.fcs) >= 1):
for f in self.fcs:
out = f(out)
out = self.relu(out)
out = self.output(out)
return out
# 1D convolutional neural network
class ConvNeuralNet(nn.Module):
def __init__(self, input_size, units, output_size):
super(ConvNeuralNet, self).__init__()
self.input = nn.Sequential(
nn.Conv1d(1, units[0], kernel_size=10),
nn.ReLU(),
nn.MaxPool1d(kernel_size=1, stride=1)
)
self.hiddens = []
if(len(units) > 1):
for i in range(0, len(units)-1):
hidden = nn.Sequential(
nn.Conv1d(units[i], units[i+1], kernel_size=10),
nn.ReLU(),
nn.MaxPool1d(kernel_size=1, stride=1)
)
self.hiddens.append(hidden)
self.units = units
def forward(self, x):
out = self.input(x)
if(len(self.hiddens) >= 1):
i = 1
for h in self.hiddens:
out = h(out)
i += 1
out_ch = out.size(2)
out = out.view(out.size(0), -1)
output = nn.Sequential(
nn.Linear(self.units[-1]*out_ch, 5)
)
out = output(out)
return out
# 1D convolutional neural network
class ConvNeuralNet2Dense(nn.Module):
def __init__(self, input_channels, units, output_size):
super(ConvNeuralNet2Dense, self).__init__()
self.input = nn.Sequential(
nn.Conv1d(input_channels, units[0], kernel_size=5, stride=1),
nn.ReLU(),
nn.MaxPool1d(kernel_size=2))
self.h1 = nn.Sequential(
nn.Conv1d(units[0], units[1], kernel_size=5,
stride=1),
nn.ReLU(),
nn.MaxPool1d(kernel_size=2))
self.fc1 = nn.Sequential(
nn.Linear(units[1]*47, units[2]),
nn.ReLU()
)
self.fc2 = nn.Sequential(
nn.Linear(units[2], units[3]),
nn.ReLU()
)
self.output = nn.Sequential(
nn.Linear(units[3], output_size)
)
self.units = units
def forward(self, x):
out = self.input(x)
out = self.h1(out)
# out_ch = out.size(2)
out = out.view(out.size(0), -1)
out = self.fc1(out)
out = self.fc2(out)
out = self.output(out)
return out