-
Notifications
You must be signed in to change notification settings - Fork 0
/
FBRNN.py
82 lines (70 loc) · 2.74 KB
/
FBRNN.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
import torch
import torch.nn as nn
import numpy as np
import scipy
def Filterbank(x, sampling, filterIdx):
#x: time signal, np array with format (electrodes,data)
#sampling: sampling frequency
#filterIdx: filter index
passband = [6, 14, 22, 30, 38, 46, 54, 62, 70, 78]
stopband = [4, 10, 16, 24, 32, 40, 48, 56, 64, 72]
Nq = sampling/2
Wp = [passband[filterIdx]/Nq, 90/Nq]
Ws = [stopband[filterIdx]/Nq, 100/Nq]
[N, Wn] = scipy.signal.cheb1ord(Wp, Ws, 3, 40)
[B, A] = scipy.signal.cheby1(N, 0.5, Wn, 'bandpass')
y = np.zeros(x.shape)
channels = x.shape[0]
for c in range(channels):
y[c, :] = scipy.signal.filtfilt(B, A, x[c, :], padtype = 'odd', padlen=3*(max(len(B),len(A))-1), axis=-1)
return y
class FBRNN(nn.Module):
def __init__(self):
super(FBRNN, self).__init__()
drop=0.4
self.conv1= nn.Conv1d(10, 8, kernel_size=(32))
self.bn1=nn.BatchNorm1d(8)
self.relu1= nn.ReLU()
self.pool1= nn.MaxPool1d(kernel_size=(2), stride=(2))
self.drop1= nn.Dropout(p=drop)
self.conv2= nn.Conv1d(8, 10, kernel_size=(32))
self.bn2=nn.BatchNorm1d(10)
self.relu2= nn.ReLU()
self.pool2= nn.MaxPool1d(kernel_size=(2), stride=(2))
self.drop2= nn.Dropout(p=drop)
self.LSTM1= nn.LSTM(input_size=10,hidden_size=100,num_layers=1,batch_first=True,bidirectional=False)
self.drop3= nn.Dropout(p=drop)
self.LSTM2= nn.LSTM(input_size=100,hidden_size=50,num_layers=1,batch_first=True,bidirectional=False)
self.drop4= nn.Dropout(p=drop)
self.LSTM3= nn.LSTM(input_size=50,hidden_size=20,num_layers=1,batch_first=True,bidirectional=False)
self.drop5= nn.Dropout(p=drop)
self.LSTM4= nn.LSTM(input_size=20,hidden_size=10,num_layers=1,batch_first=True,bidirectional=False)
self.drop6= nn.Dropout(p=drop)
self.LSTM5= nn.LSTM(input_size=10,hidden_size=5,num_layers=1,batch_first=True,bidirectional=False)
self.drop7= nn.Dropout(p=drop)
self.outL= nn.Linear(40,len(targets))
def forward(self,x):
x=self.conv1(x)
x=self.bn1(x)
x=self.relu1(x)
x=self.pool1(x)
x=self.drop1(x)
x=self.conv2(x)
x=self.bn2(x)
x=self.relu2(x)
x=self.pool2(x)
x=self.drop2(x)
x=x.permute(0,2,1)#batch,sequence,channels
x,_=self.LSTM1(x)
x=self.drop3(x)
x,_=self.LSTM2(x)
x=self.drop4(x)
x,_=self.LSTM3(x)
x=self.drop5(x)
x,_=self.LSTM4(x)
x=self.drop6(x)
x,_=self.LSTM5(x)
x=self.drop7(x)
x=x.reshape(x.size(0), -1)
x=self.outL(x)
return(x)