File size: 9,467 Bytes
0164e4a
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
import torch
from torch import nn
from torch.nn import functional as F
import modules

from torch.nn import Conv1d, ConvTranspose1d, AvgPool1d, Conv2d
from torch.nn.utils import weight_norm, remove_weight_norm, spectral_norm
from commons import init_weights, get_padding
from torch.cuda.amp import autocast
import torchaudio
from einops import rearrange

from alias_free_torch import *
import activations

class AMPBlock0(torch.nn.Module):
    def __init__(self, channels, kernel_size=3, dilation=(1, 3, 5), activation=None):
        super(AMPBlock0, self).__init__()
      
        self.convs1 = nn.ModuleList([
            weight_norm(Conv1d(channels, channels, kernel_size, 1, dilation=dilation[0],
                               padding=get_padding(kernel_size, dilation[0]))),
            weight_norm(Conv1d(channels, channels, kernel_size, 1, dilation=dilation[1],
                               padding=get_padding(kernel_size, dilation[1]))),
            weight_norm(Conv1d(channels, channels, kernel_size, 1, dilation=dilation[2],
                               padding=get_padding(kernel_size, dilation[2]))),
        ])
        self.convs1.apply(init_weights)

        self.convs2 = nn.ModuleList([
            weight_norm(Conv1d(channels, channels, kernel_size, 1, dilation=1,
                               padding=get_padding(kernel_size, 1))),
            weight_norm(Conv1d(channels, channels, kernel_size, 1, dilation=1,
                               padding=get_padding(kernel_size, 1))),
            weight_norm(Conv1d(channels, channels, kernel_size, 1, dilation=1,
                               padding=get_padding(kernel_size, 1))),
        ])
        self.convs2.apply(init_weights)

        self.num_layers = len(self.convs1) + len(self.convs2) # total number of conv layers


        self.activations = nn.ModuleList([
            Activation1d(
                activation=activations.SnakeBeta(channels, alpha_logscale=True))
                for _ in range(self.num_layers)
        ])
  
    def forward(self, x):
        acts1, acts2 = self.activations[::2], self.activations[1::2]
        for c1, c2, a1, a2 in zip(self.convs1, self.convs2, acts1, acts2):
            xt = a1(x)
            xt = c1(xt)
            xt = a2(xt)
            xt = c2(xt)
            x = xt + x

        return x

    def remove_weight_norm(self):
        for l in self.convs1:
            remove_weight_norm(l)
        for l in self.convs2:
            remove_weight_norm(l)


class Generator(torch.nn.Module):
    def __init__(self, initial_channel, resblock, resblock_kernel_sizes, resblock_dilation_sizes, upsample_rates, upsample_initial_channel, upsample_kernel_sizes, gin_channels=0):
        super(Generator, self).__init__()
        self.num_kernels = len(resblock_kernel_sizes)
        self.num_upsamples = len(upsample_rates)
        
        self.conv_pre = weight_norm(Conv1d(initial_channel, upsample_initial_channel, 7, 1, padding=3))
        resblock = AMPBlock0

        self.resblocks = nn.ModuleList()
        for i in range(1):
            ch = upsample_initial_channel//(2**(i))
            for j, (k, d) in enumerate(zip(resblock_kernel_sizes, resblock_dilation_sizes)):
                self.resblocks.append(resblock(ch, k, d, activation="snakebeta"))

        activation_post = activations.SnakeBeta(ch, alpha_logscale=True)
        self.activation_post = Activation1d(activation=activation_post)

        self.conv_post = Conv1d(ch, 1, 7, 1, padding=3, bias=False)
        if gin_channels != 0:
            self.cond = nn.Conv1d(gin_channels, upsample_initial_channel, 1)

    def forward(self, x, g=None):
        x = self.conv_pre(x)
        if g is not None:
          x = x + self.cond(g)

        for i in range(self.num_upsamples):
      
            x = F.interpolate(x, int(x.shape[-1] * 1.5), mode='linear')
            xs = None
            for j in range(self.num_kernels):
                if xs is None:
                    xs = self.resblocks[i*self.num_kernels+j](x)
                else:
                    xs += self.resblocks[i*self.num_kernels+j](x)
            x = xs / self.num_kernels

        x = self.activation_post(x)
        x = self.conv_post(x)
        x = torch.tanh(x)

        return x

    def remove_weight_norm(self):
        print('Removing weight norm...')
        for l in self.resblocks:
            l.remove_weight_norm()
        remove_weight_norm(self.conv_pre)

class DiscriminatorP(torch.nn.Module):
    def __init__(self, period, kernel_size=5, stride=3, use_spectral_norm=False):
        super(DiscriminatorP, self).__init__()
        self.period = period
        self.use_spectral_norm = use_spectral_norm
        norm_f = weight_norm if use_spectral_norm == False else spectral_norm
        self.convs = nn.ModuleList([
            norm_f(Conv2d(1, 32, (kernel_size, 1), (stride, 1), padding=(get_padding(kernel_size, 1), 0))),
            norm_f(Conv2d(32, 128, (kernel_size, 1), (stride, 1), padding=(get_padding(kernel_size, 1), 0))),
            norm_f(Conv2d(128, 512, (kernel_size, 1), (stride, 1), padding=(get_padding(kernel_size, 1), 0))),
            norm_f(Conv2d(512, 1024, (kernel_size, 1), (stride, 1), padding=(get_padding(kernel_size, 1), 0))),
            norm_f(Conv2d(1024, 1024, (kernel_size, 1), 1, padding=(get_padding(kernel_size, 1), 0))),
        ])
        self.conv_post = norm_f(Conv2d(1024, 1, (3, 1), 1, padding=(1, 0)))

    def forward(self, x):
        fmap = []

        # 1d to 2d
        b, c, t = x.shape
        if t % self.period != 0: # pad first
            n_pad = self.period - (t % self.period)
            x = F.pad(x, (0, n_pad), "reflect")
            t = t + n_pad
        x = x.view(b, c, t // self.period, self.period)

        for l in self.convs:
            x = l(x)
            x = F.leaky_relu(x, modules.LRELU_SLOPE)
            fmap.append(x)
        x = self.conv_post(x)
        fmap.append(x)
        x = torch.flatten(x, 1, -1)

        return x, fmap

class DiscriminatorR(torch.nn.Module):
    def __init__(self, resolution, use_spectral_norm=False):
        super(DiscriminatorR, self).__init__()
        norm_f = weight_norm if use_spectral_norm == False else spectral_norm

        n_fft, hop_length, win_length = resolution
        self.spec_transform = torchaudio.transforms.Spectrogram(
            n_fft=n_fft, hop_length=hop_length, win_length=win_length, window_fn=torch.hann_window,
            normalized=True, center=False, pad_mode=None, power=None)

        self.convs = nn.ModuleList([
            norm_f(nn.Conv2d(2, 32, (3, 9), padding=(1, 4))),
            norm_f(nn.Conv2d(32, 32, (3, 9), stride=(1, 2), padding=(1, 4))),
            norm_f(nn.Conv2d(32, 32, (3, 9), stride=(1, 2), dilation=(2,1), padding=(2, 4))),
            norm_f(nn.Conv2d(32, 32, (3, 9), stride=(1, 2), dilation=(4,1), padding=(4, 4))),
            norm_f(nn.Conv2d(32, 32, (3, 3), padding=(1, 1))),
        ])
        self.conv_post = norm_f(nn.Conv2d(32, 1, (3, 3), padding=(1, 1)))

    def forward(self, y):
        fmap = []

        x = self.spec_transform(y)  # [B, 2, Freq, Frames, 2]
        x = torch.cat([x.real, x.imag], dim=1)
        x = rearrange(x, 'b c w t -> b c t w')

        for l in self.convs:
            x = l(x)
            x = F.leaky_relu(x, modules.LRELU_SLOPE)
            fmap.append(x)
        x = self.conv_post(x)
        fmap.append(x)
        x = torch.flatten(x, 1, -1)

        return x, fmap


class MultiPeriodDiscriminator(torch.nn.Module):
    def __init__(self, use_spectral_norm=False):
        super(MultiPeriodDiscriminator, self).__init__()
        periods = [2,3,5,7,11]
        resolutions = [[2048, 512, 2048], [1024, 256, 1024], [512, 128, 512], [256, 64, 256], [128, 32, 128]]

        discs = [DiscriminatorR(resolutions[i], use_spectral_norm=use_spectral_norm) for i in range(len(resolutions))]
        discs = discs + [DiscriminatorP(i, use_spectral_norm=use_spectral_norm) for i in periods]
        self.discriminators = nn.ModuleList(discs)

    def forward(self, y, y_hat):
        y_d_rs = []
        y_d_gs = []
        fmap_rs = []
        fmap_gs = []
        for i, d in enumerate(self.discriminators):
            y_d_r, fmap_r = d(y)
            y_d_g, fmap_g = d(y_hat)
            y_d_rs.append(y_d_r)
            y_d_gs.append(y_d_g)
            fmap_rs.append(fmap_r)
            fmap_gs.append(fmap_g)

        return y_d_rs, y_d_gs, fmap_rs, fmap_gs

class SynthesizerTrn(nn.Module):
  """
  Synthesizer for Training
  """

  def __init__(self,

    spec_channels,
    segment_size,
    resblock,
    resblock_kernel_sizes,
    resblock_dilation_sizes,
    upsample_rates,
    upsample_initial_channel,
    upsample_kernel_sizes,
    **kwargs):

    super().__init__()
    self.spec_channels = spec_channels
    self.resblock = resblock
    self.resblock_kernel_sizes = resblock_kernel_sizes
    self.resblock_dilation_sizes = resblock_dilation_sizes
    self.upsample_rates = upsample_rates
    self.upsample_initial_channel = upsample_initial_channel
    self.upsample_kernel_sizes = upsample_kernel_sizes
    self.segment_size = segment_size

    self.dec = Generator(1, resblock, resblock_kernel_sizes, resblock_dilation_sizes, upsample_rates, upsample_initial_channel, upsample_kernel_sizes)

  def forward(self, x):

    y = self.dec(x)
    return y
  @torch.no_grad()
  def infer(self, x, max_len=None):

    o = self.dec(x[:,:,:max_len])
    return o