from typing import List, Tuple import torch from torch import Tensor, nn from torch.nn import Conv1d, ConvTranspose1d, Module import torch.nn.functional as F from torch.nn.utils import remove_weight_norm, weight_norm from models.config import HifiGanConfig, HifiGanPretrainingConfig, PreprocessingConfig from .utils import get_padding, init_weights # Leaky ReLU slope LRELU_SLOPE = HifiGanPretrainingConfig.lReLU_slope class ResBlock1(Module): def __init__( self, channels: int, kernel_size: int = 3, dilation: List[int] = [1, 3, 5], ): r"""Initialize the ResBlock1 module. Args: channels (int): The number of channels for the ResBlock. kernel_size (int, optional): The kernel size for the convolutional layers. Defaults to 3. dilation (Tuple[int, int, int], optional): The dilation for the convolutional layers. Defaults to (1, 3, 5). """ super().__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) def forward(self, x: Tensor) -> Tensor: r"""Forward pass of the ResBlock1 module. Args: x (Tensor): The input tensor. Returns: Tensor: The output tensor. """ for c1, c2 in zip(self.convs1, self.convs2): xt = F.leaky_relu(x, LRELU_SLOPE) xt = c1(xt) xt = F.leaky_relu(xt, LRELU_SLOPE) xt = c2(xt) x = xt + x return x def remove_weight_norm(self): r"""Remove the weight normalization from the convolutional layers.""" for layer in self.convs1: remove_weight_norm(layer) for layer in self.convs2: remove_weight_norm(layer) class ResBlock2(Module): def __init__( self, channels: int, kernel_size: int = 3, dilation: List[int] = [1, 3], ): r"""Initialize the ResBlock2 module. Args: channels (int): The number of channels for the ResBlock. kernel_size (int, optional): The kernel size for the convolutional layers. Defaults to 3. dilation (Tuple[int, int], optional): The dilation for the convolutional layers. Defaults to (1, 3). """ super().__init__() self.convs = 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]), ), ), ], ) self.convs.apply(init_weights) def forward(self, x: Tensor) -> Tensor: r"""Forward pass of the ResBlock2 module. Args: x (Tensor): The input tensor. Returns: Tensor: The output tensor. """ for layer in self.convs: xt = F.leaky_relu(x, LRELU_SLOPE) xt = layer(xt) x = xt + x return x def remove_weight_norm(self): r"""Remove the weight normalization from the convolutional layers.""" for layer in self.convs: remove_weight_norm(layer) class Generator(Module): def __init__(self, h: HifiGanConfig, p: PreprocessingConfig): r"""Initialize the Generator module. Args: h (HifiGanConfig): The configuration for the Generator. p (PreprocessingConfig): The configuration for the preprocessing. """ super().__init__() self.h = h self.p = p self.num_kernels = len(h.resblock_kernel_sizes) self.num_upsamples = len(h.upsample_rates) self.conv_pre = weight_norm( Conv1d( self.p.stft.n_mel_channels, h.upsample_initial_channel, 7, 1, padding=3, ), ) resblock = ResBlock1 if h.resblock == "1" else ResBlock2 self.ups = nn.ModuleList() for i, (u, k) in enumerate(zip(h.upsample_rates, h.upsample_kernel_sizes)): self.ups.append( weight_norm( ConvTranspose1d( h.upsample_initial_channel // (2**i), h.upsample_initial_channel // (2 ** (i + 1)), k, u, padding=(k - u) // 2, ), ), ) self.resblocks = nn.ModuleList() for i in range(len(self.ups)): resblock_list = nn.ModuleList() ch = h.upsample_initial_channel // (2 ** (i + 1)) for _, (k, d) in enumerate( zip(h.resblock_kernel_sizes, h.resblock_dilation_sizes), ): resblock_list.append(resblock(ch, k, d)) self.resblocks.append(resblock_list) self.conv_post = weight_norm(Conv1d(ch, 1, 7, 1, padding=3)) self.ups.apply(init_weights) self.conv_post.apply(init_weights) def forward(self, x: Tensor) -> Tensor: r"""Forward pass of the Generator module. Args: x (Tensor): The input tensor. Returns: Tensor: The output tensor. """ x = self.conv_pre(x) for upsample_layer, resblock_group in zip(self.ups, self.resblocks): x = F.leaky_relu(x, LRELU_SLOPE) x = upsample_layer(x) xs = torch.zeros(x.shape, dtype=x.dtype, device=x.device) for resblock in resblock_group: # type: ignore xs += resblock(x) x = xs / self.num_kernels x = F.leaky_relu(x) x = self.conv_post(x) x = torch.tanh(x) return x