-
Notifications
You must be signed in to change notification settings - Fork 2
/
Copy pathmodels.py
184 lines (159 loc) · 6.94 KB
/
models.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
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
180
181
182
'''
pytorch implementation Adrien Bitton
link: https://github.com/adrienchaton/PerceptualAudio_pytorch
paper codes Pranay Manocha
link: https://github.com/pranaymanocha/PerceptualAudio
'''
import torch
import torch.nn as nn
import torch.nn.functional as F
###############################################################################
### sub networks
class lossnet(nn.Module):
def __init__(self,nconv=14,nchan=32,dp=0.1,dist_act='no'):
# base settings for 16kHz, applied to 22kHz
# in the case of dropout, at training, forward over two same tensors does not give dist=0
# the droupout is randomized differently for each pass on xref/xper
super(lossnet, self).__init__()
self.nconv = nconv
self.dist_act = dist_act
self.convs = nn.ModuleList()
self.chan_w = nn.ParameterList()
for iconv in range(nconv):
if iconv==0:
chin = 1
else:
chin = nchan
if (iconv+1)%5==0:
nchan = nchan*2
if iconv<nconv-1:
conv = [nn.Conv1d(chin,nchan,3,stride=2,padding=1),nn.BatchNorm1d(nchan),nn.LeakyReLU()]
if dp!=0:
conv.append(nn.Dropout(p=dp))
else:
# last conv has no stride and no dropout
conv = [nn.Conv1d(chin,nchan,3,stride=1,padding=1),nn.BatchNorm1d(nchan),nn.LeakyReLU()]
self.convs.append(nn.Sequential(*conv))
self.chan_w.append(nn.Parameter(torch.randn(nchan),requires_grad=True))
if dist_act=='sig':
self.act = nn.Sigmoid()
elif dist_act=='tanh':
self.act = nn.Tanh()
elif dist_act=='tshrink':
self.act = nn.Tanhshrink()
elif dist_act=='exp':
self.act = None
elif dist_act=='no':
self.act = nn.Identity()
else:
self.act = None
def forward(self,xref,xper):
# xref and xper are [batch,L]
xref = xref.unsqueeze(1)
xper = xper.unsqueeze(1)
dist = 0
for iconv in range(self.nconv):
xref = self.convs[iconv](xref)
xper = self.convs[iconv](xper)
diff = (xref-xper).permute(0,2,1) # channel last
wdiff = diff*self.chan_w[iconv]
wdiff = torch.sum(torch.abs(wdiff),dim=(1,2))/diff.shape[1]/diff.shape[2] # average by time and channel dimensions
dist = dist+wdiff
if self.dist_act=='exp':
dist = torch.exp(torch.clamp(dist,max=20.))/(10**5) # exp(20) ~ 4*10**8
else:
dist = self.act(dist)
return dist
class classifnet(nn.Module):
def __init__(self,ndim=[16,6],dp=0.1,BN=1,classif_act='no'):
# lossnet is pair of [batch,L] -> dist [batch]
# classifnet goes dist [batch] -> pred [batch,2] == evaluate BCE with low-capacity
super(classifnet, self).__init__()
n_layers = 2
MLP = []
for ilayer in range(n_layers):
if ilayer==0:
fin = 1
else:
fin = ndim[ilayer-1]
MLP.append(nn.Linear(fin,ndim[ilayer]))
if BN==1 and ilayer==0: # only 1st hidden layer
MLP.append(nn.BatchNorm1d(ndim[ilayer]))
elif BN==2: # the two hidden layers
MLP.append(nn.BatchNorm1d(ndim[ilayer]))
MLP.append(nn.LeakyReLU())
if dp!=0:
MLP.append(nn.Dropout(p=dp))
# last linear maps to binary class probabilities ; loss includes LogSoftmax
MLP.append(nn.Linear(ndim[ilayer],2))
if classif_act=='sig':
MLP.append(nn.Sigmoid())
if classif_act=='tanh':
MLP.append(nn.Tanh())
self.MLP = nn.Sequential(*MLP)
def forward(self,dist):
return self.MLP(dist.unsqueeze(1))
###############################################################################
### full model
def weights_init(m):
classname = m.__class__.__name__
if classname.find('Conv') != -1 or classname.find('BatchNorm') != -1 or classname.find('Linear') != -1:
torch.nn.init.normal_(m.weight)
# default Linear init is kaiming_uniform_ / default Conv1d init is a scaled uniform / default BN init is constant gamma=1 and bias=0
try:
torch.nn.init.constant_(m.bias, 0.01)
except:
pass
# else:
# # they are only non-trainable classes eg. Relu,Dropout,Sequential ...
# print(classname)
class JNDnet(nn.Module):
def __init__(self,nconv=14,nchan=32,dist_dp=0.1,dist_act='no',ndim=[16,6],classif_dp=0.1,classif_BN=0,classif_act='no',dev=torch.device('cpu'),minit=0):
super(JNDnet, self).__init__()
self.model_dist = lossnet(nconv=nconv,nchan=nchan,dp=dist_dp,dist_act=dist_act)
self.model_classif = classifnet(ndim=ndim,dp=classif_dp,BN=classif_BN,classif_act=classif_act)
if minit==1:
self.model_dist.apply(weights_init) # custom weight initialization
self.model_classif.apply(weights_init)
self.CE = nn.CrossEntropyLoss(reduction='mean')
self.dev = dev
def forward(self,xref,xper,labels):
dist = self.model_dist.forward(xref,xper)
pred = self.model_classif.forward(dist)
# loss = self.CE(pred,labels.squeeze(1)) # pred is [batch,2] and labels [batch] long and binary
loss = self.CE(pred,torch.squeeze(labels,-1))
class_prob = F.softmax(pred,dim=-1)
class_pred = torch.argmax(class_prob,dim=-1)
return loss,dist,class_pred,class_prob
def grad_check(self,minibatch,optimizer):
xref = minibatch[0].to(self.dev)
xper = minibatch[1].to(self.dev)
labels = minibatch[2].to(self.dev)
loss,dist,class_pred,class_prob = self.forward(xref,xper,labels)
print('\nbackward on classification loss')
optimizer.zero_grad()
loss.backward()
tot_grad = 0
for name, param in self.named_parameters():
if param.grad is not None:
sum_abs_paramgrad = torch.sum(torch.abs(param.grad)).item()
if sum_abs_paramgrad==0:
print(name,'sum_abs_paramgrad==0')
else:
tot_grad += sum_abs_paramgrad
else:
print(name,'param.grad is None')
print('tot_grad = ',tot_grad)
norm_type = 2
loss,dist,class_pred,class_prob = self.forward(xref,xper,labels)
optimizer.zero_grad()
loss.backward()
total_norm = 0
for name, param in self.named_parameters():
if param.grad is not None:
param_norm = param.grad.data.norm(norm_type)
total_norm += param_norm.item() ** norm_type
else:
print(name,'param.grad is None')
total_norm = total_norm ** (1. / norm_type)
print('total_norm over all layers ==',total_norm)