File size: 4,799 Bytes
9d61c9b
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
from typing import Any, Tuple

import torch
from torch import nn
from torch.nn import Module
import torch.nn.functional as F
from torch.nn.utils.parametrizations import spectral_norm, weight_norm

from models.config import VocoderModelConfig


class DiscriminatorR(Module):
    r"""A class representing the Residual Discriminator network for a UnivNet vocoder.

    Args:
        resolution (Tuple): A tuple containing the number of FFT points, hop length, and window length.
        model_config (VocoderModelConfig): A configuration object for the UnivNet model.
    """

    def __init__(
        self,
        resolution: Tuple[int, int, int],
        model_config: VocoderModelConfig,
    ):
        super().__init__()

        self.resolution = resolution
        self.LRELU_SLOPE = model_config.mrd.lReLU_slope

        # Use spectral normalization or weight normalization based on the configuration
        norm_f: Any = (
            spectral_norm if model_config.mrd.use_spectral_norm else weight_norm
        )

        # Define the convolutional layers
        self.convs = nn.ModuleList(
            [
                norm_f(
                    nn.Conv2d(
                        1,
                        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),
                        padding=(1, 4),
                    ),
                ),
                norm_f(
                    nn.Conv2d(
                        32,
                        32,
                        (3, 9),
                        stride=(1, 2),
                        padding=(1, 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, x: torch.Tensor) -> tuple[list[torch.Tensor], torch.Tensor]:
        r"""Forward pass of the DiscriminatorR class.

        Args:
            x (torch.Tensor): The input tensor.

        Returns:
            tuple: A tuple containing the intermediate feature maps and the output tensor.
        """
        fmap = []

        # Compute the magnitude spectrogram of the input waveform
        x = self.spectrogram(x)

        # Add a channel dimension to the spectrogram tensor
        x = x.unsqueeze(1)

        # Apply the convolutional layers with leaky ReLU activation
        for layer in self.convs:
            x = layer(x.to(dtype=self.conv_post.weight.dtype))
            x = F.leaky_relu(x, self.LRELU_SLOPE)
            fmap.append(x)

        # Apply the post-convolutional layer
        x = self.conv_post(x)
        fmap.append(x)

        # Flatten the output tensor
        x = torch.flatten(x, 1, -1)

        return fmap, x

    def spectrogram(self, x: torch.Tensor) -> torch.Tensor:
        r"""Computes the magnitude spectrogram of the input waveform.

        Args:
            x (torch.Tensor): Input waveform tensor of shape [B, C, T].

        Returns:
            torch.Tensor: Magnitude spectrogram tensor of shape [B, F, TT], where F is the number of frequency bins and TT is the number of time frames.
        """
        n_fft, hop_length, win_length = self.resolution

        # Apply reflection padding to the input waveform
        x = F.pad(
            x,
            (int((n_fft - hop_length) / 2), int((n_fft - hop_length) / 2)),
            mode="reflect",
        )

        # Squeeze the input waveform to remove the channel dimension
        x = x.squeeze(1)

        # Compute the short-time Fourier transform of the input waveform
        x = torch.stft(
            x,
            n_fft=n_fft,
            hop_length=hop_length,
            win_length=win_length,
            center=False,
            return_complex=True,
            window=torch.ones(win_length, device=x.device),
        )  # [B, F, TT, 2]

        x = torch.view_as_real(x)

        # Compute the magnitude spectrogram from the complex spectrogram
        return torch.norm(x, p=2, dim=-1)  # [B, F, TT]