-
Notifications
You must be signed in to change notification settings - Fork 6
/
repgan.py
181 lines (141 loc) · 5.49 KB
/
repgan.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
import torch
import torch.nn as nn
from rep_conv import RepConv
class Upsample(nn.Module):
def __init__(self,
in_channels,
out_channels,
scale=2):
super(Upsample, self).__init__()
self.scale = scale
self.transposed_conv = nn.ConvTranspose1d(in_channels,
out_channels,
kernel_size=scale * 2,
stride=scale,
padding=scale // 2 + scale % 2,
output_padding=scale % 2)
def forward(self, x):
return self.transposed_conv(x)
def inference(self, x):
return self.forward(x)
class ResidualBlock(nn.Module):
def __init__(self,
channels: int = 512,
kernel_sizes: tuple = (3, 7, 11),
dilations: tuple = (1, 3, 5),
use_additional_convs: bool = True):
super(ResidualBlock, self).__init__()
self.use_additional_convs = use_additional_convs
self.act = nn.LeakyReLU(0.1)
self.convs1 = nn.ModuleList()
if use_additional_convs:
self.convs2 = nn.ModuleList()
for dilation in dilations:
self.convs1.append(RepConv(channels, kernel_sizes, dilation=dilation))
if use_additional_convs:
self.convs2 += [RepConv(channels, kernel_sizes, dilation=1)]
def forward(self, x: torch.Tensor) -> torch.Tensor:
for idx in range(len(self.convs1)):
x = self.act(x)
x = self.convs1[idx](x)
if self.use_additional_convs:
x = self.act(x)
x = self.convs2[idx](x)
return x
def inference(self, x: torch.Tensor) -> torch.Tensor:
for idx in range(len(self.convs1)):
x = self.act(x)
x = self.convs1[idx].inference(x)
if self.use_additional_convs:
x = self.act(x)
x = self.convs2[idx].inference(x)
return x
class RepGANGenerator(nn.Module):
def __init__(self,
in_channels=80,
out_channels=1,
channels=512,
kernel_size=7,
dropout=0.1,
upsample_scales=(8, 8, 2, 2),
resblock_kernel_sizes=(3, 7, 11),
resblock_dilations=((1, 3, 5), (1, 3, 5), (1, 3, 5), (1, 3, 5)),
use_additional_convs=True,
use_weight_norm=True,
):
super(RepGANGenerator, self).__init__()
# check hyper parameters are valid
assert kernel_size % 2 == 1, "Kernel size must be odd number."
assert len(resblock_dilations) == len(upsample_scales)
self.input_conv = nn.Sequential(
nn.Conv1d(in_channels, channels, kernel_size=kernel_size, padding=kernel_size // 2),
nn.LeakyReLU(0.1),
nn.Dropout(dropout)
)
self.upsamples = nn.ModuleList()
self.blocks = nn.ModuleList()
self.non_linear = nn.LeakyReLU(0.1)
for i in range(len(upsample_scales)):
self.upsamples += [
Upsample(
in_channels=channels // (2 ** i),
out_channels=channels // (2 ** (i + 1)),
scale=upsample_scales[i]
)
]
self.blocks += [
ResidualBlock(
channels=channels // (2 ** (i + 1)),
kernel_sizes=resblock_kernel_sizes,
dilations=resblock_dilations[i],
use_additional_convs=use_additional_convs
)
]
self.output_conv = nn.Sequential(
nn.Conv1d(channels // (2 ** (i + 1)), out_channels, kernel_size, padding=kernel_size // 2),
nn.Tanh()
)
if use_weight_norm:
self.apply_weight_norm()
self.reset_parameters()
def forward(self, x):
x = self.input_conv(x)
for i in range(len(self.upsamples)):
x = self.upsamples[i](x)
if self.training:
x = self.blocks[i](x)
else:
x = self.blocks[i].inference(x)
x = self.non_linear(x)
x = self.output_conv(x)
return x
def convert_weight_bias(self):
def _convert_weight_bias(m):
if isinstance(m, RepConv):
m.convert_weight_bias()
self.apply(_convert_weight_bias)
def apply_weight_norm(self):
"""Apply weight normalization module from all the layers."""
def _apply_weight_norm(m):
if isinstance(m, nn.Conv1d):
nn.utils.weight_norm(m)
if isinstance(m, nn.ConvTranspose1d):
nn.utils.weight_norm(m, dim=1)
self.apply(_apply_weight_norm)
def reset_parameters(self):
def _reset_parameters(m):
if isinstance(m, (nn.Conv1d, nn.ConvTranspose1d)):
m.weight.data.normal_(0.0, 0.02)
self.apply(_reset_parameters)
if __name__ == '__main__':
voc = RepGANGenerator(use_weight_norm=False)
dummy = torch.rand(1, 80, 100)
y1 = voc(dummy)
params = 0
for n, p in voc.named_parameters():
params += p.numel()
print('model size: {}M'.format(params / 1e6))
voc.convert_weight_bias()
voc.eval()
y2 = voc(dummy)
print((y1 - y2).mean())