def __init__(self, hparams):
super(Postnet, self).__init__()
self.convolutions = nn.ModuleList()
self.convolutions.append(
nn.Sequential(
ConvNorm(hparams.n_acoustic_feat_dims,
hparams.postnet_embedding_dim,
kernel_size=hparams.postnet_kernel_size, stride=1,
padding=int((hparams.postnet_kernel_size - 1) / 2),
dilation=1, w_init_gain='tanh'),
nn.BatchNorm1d(hparams.postnet_embedding_dim))
)
for i in range(1, hparams.postnet_n_convolutions - 1):
self.convolutions.append(
nn.Sequential(
ConvNorm(hparams.postnet_embedding_dim,
hparams.postnet_embedding_dim,
kernel_size=hparams.postnet_kernel_size, stride=1,
padding=int((hparams.postnet_kernel_size - 1) / 2),
dilation=1, w_init_gain='tanh'),
nn.BatchNorm1d(hparams.postnet_embedding_dim))
)
self.convolutions.append(
nn.Sequential(
ConvNorm(hparams.postnet_embedding_dim,
hparams.n_acoustic_feat_dims,
kernel_size=hparams.postnet_kernel_size, stride=1,
padding=int((hparams.postnet_kernel_size - 1) / 2),
dilation=1, w_init_gain='linear'),
nn.BatchNorm1d(hparams.n_acoustic_feat_dims))
)
def __init__(self, n_mel_channels, postnet_embedding_dim,
postnet_kernel_size, postnet_n_convolutions):
super(Postnet, self).__init__()
self.convolutions = nn.ModuleList()
self.convolutions.append(
nn.Sequential(
ConvNorm(n_mel_channels, postnet_embedding_dim,
kernel_size=postnet_kernel_size, stride=1,
padding=int((postnet_kernel_size - 1) / 2),
dilation=1, w_init_gain='tanh'),
nn.BatchNorm1d(postnet_embedding_dim))
)
for i in range(1, postnet_n_convolutions - 1):
self.convolutions.append(
nn.Sequential(
ConvNorm(postnet_embedding_dim,
postnet_embedding_dim,
kernel_size=postnet_kernel_size, stride=1,
padding=int((postnet_kernel_size - 1) / 2),
dilation=1, w_init_gain='tanh'),
nn.BatchNorm1d(postnet_embedding_dim))
)
self.convolutions.append(
nn.Sequential(
ConvNorm(postnet_embedding_dim, n_mel_channels,
kernel_size=postnet_kernel_size, stride=1,
padding=int((postnet_kernel_size - 1) / 2),
dilation=1, w_init_gain='linear'),
nn.BatchNorm1d(n_mel_channels))
)
self.n_convs = len(self.convolutions)
def __init__(self, hparams):
super(Encoder, self).__init__()
self.prenet = Prenet(
hparams.n_symbols,
[hparams.symbols_embedding_dim, hparams.symbols_embedding_dim])
convolutions = []
for _ in range(hparams.encoder_n_convolutions):
conv_layer = nn.Sequential(
ConvNorm(hparams.encoder_embedding_dim,
hparams.encoder_embedding_dim,
kernel_size=hparams.encoder_kernel_size,
stride=1,
padding=int((hparams.encoder_kernel_size - 1) / 2),
dilation=1,
w_init_gain='relu'),
nn.BatchNorm1d(hparams.encoder_embedding_dim))
convolutions.append(conv_layer)
self.convolutions = nn.ModuleList(convolutions)
self.lstm = nn.LSTM(hparams.encoder_embedding_dim,
int(hparams.encoder_embedding_dim / 2),
1,
batch_first=True,
bidirectional=True)
def __init__(
self,
encoder_n_convolutions, # Number of encoder convolutions; def. 3
encoder_embedding_dim, # Encoder embedding dimension; def. 512
encoder_kernel_size # Encoder kernel size; def. 5
):
super(Encoder, self).__init__()
convolutions = []
for _ in range(encoder_n_convolutions):
conv_layer = nn.Sequential(
ConvNorm(encoder_embedding_dim,
encoder_embedding_dim,
kernel_size=encoder_kernel_size,
stride=1,
padding=int((encoder_kernel_size - 1) / 2),
dilation=1,
w_init_gain='relu'),
nn.BatchNorm1d(encoder_embedding_dim))
convolutions.append(conv_layer)
self.convolutions = nn.ModuleList(convolutions)
self.lstm = nn.LSTM(encoder_embedding_dim,
int(encoder_embedding_dim / 2),
1,
batch_first=True,
bidirectional=True)
def __init__(self, attention_n_filters, attention_kernel_size, attention_dim):
super(LocationLayer, self).__init__()
self.location_conv = ConvNorm(1, attention_n_filters,
kernel_size=attention_kernel_size,
padding=int((attention_kernel_size - 1) / 2),
stride=1, dilation=1)
self.location_dense = LinearNorm(attention_n_filters, attention_dim,
bias=False, w_init_gain='tanh')
def __init__(
self,
n_mel_channels: int, # Number of bins in mel-spectrograms; def. 80
postnet_embedding_dim: int, # Postnet embedding dimension; def. 512
postnet_kernel_size: int, # Postnet kernel size; def. 5
postnet_n_convolutions: int # Number of postnet convolutions; def. 5
):
super(Postnet, self).__init__()
self.convolutions = nn.ModuleList()
self.convolutions.append(
nn.Sequential(
ConvNorm(n_mel_channels,
postnet_embedding_dim,
kernel_size=postnet_kernel_size,
stride=1,
padding=int((postnet_kernel_size - 1) / 2),
dilation=1,
w_init_gain='tanh'),
nn.BatchNorm1d(postnet_embedding_dim)))
for i in range(1, postnet_n_convolutions - 1):
self.convolutions.append(
nn.Sequential(
ConvNorm(postnet_embedding_dim,
postnet_embedding_dim,
kernel_size=postnet_kernel_size,
stride=1,
padding=int((postnet_kernel_size - 1) / 2),
dilation=1,
w_init_gain='tanh'),
nn.BatchNorm1d(postnet_embedding_dim)))
self.convolutions.append(
nn.Sequential(
ConvNorm(postnet_embedding_dim,
n_mel_channels,
kernel_size=postnet_kernel_size,
stride=1,
padding=int((postnet_kernel_size - 1) / 2),
dilation=1,
w_init_gain='linear'),
nn.BatchNorm1d(n_mel_channels)))
self.n_convs = len(self.convolutions)
