tuna2134/sbv2-v2
1
0
Fork 0
Code Issues Pull Requests 1 Actions Packages Projects Releases Wiki Activity

Update models.py

Browse Source
This commit is contained in:
Stardust·减
2023-08-23 17:05:18 +08:00
committed by GitHub
parent 402316fe03
commit 083500ac06
1 changed files with 5 additions and 7 deletions
Download Patch File Download Diff File Expand all files Collapse all files

12
models.py
View File

@@ -280,12 +280,12 @@ class TextEncoder(nn.Module):
gin_channels=self.gin_channels)
self.proj = nn.Conv1d(hidden_channels, out_channels * 2, 1)
def forward(self, x, x_lengths, tone, language, bert):
def forward(self, x, x_lengths, tone, language, bert, g=None):
x = (self.emb(x)+ self.tone_emb(tone)+ self.language_emb(language)+self.bert_proj(bert).transpose(1,2)) * math.sqrt(self.hidden_channels) # [b, t, h]
x = torch.transpose(x, 1, -1) # [b, h, t]
x_mask = torch.unsqueeze(commons.sequence_mask(x_lengths, x.size(2)), 1).to(x.dtype)
x = self.encoder(x * x_mask, x_mask)
x = self.encoder(x * x_mask, x_mask, g=g)
stats = self.proj(x) * x_mask
m, logs = torch.split(stats, self.out_channels, dim=1)
@@ -640,13 +640,11 @@ class SynthesizerTrn(nn.Module):
self.ref_enc = ReferenceEncoder(spec_channels, gin_channels)
def forward(self, x, x_lengths, y, y_lengths, sid, tone, language, bert):
x, m_p, logs_p, x_mask = self.enc_p(x, x_lengths, tone, language, bert)
if self.n_speakers > 0:
g = self.emb_g(sid).unsqueeze(-1) # [b, h, 1]
else:
g = self.ref_enc(y.transpose(1,2)).unsqueeze(-1)
x, m_p, logs_p, x_mask = self.enc_p(x, x_lengths, tone, language, bert,g=g)
z, m_q, logs_q, y_mask = self.enc_q(y, y_lengths, g=g)
z_p = self.flow(z, y_mask, g=g)
@@ -686,13 +684,13 @@ class SynthesizerTrn(nn.Module):
return o, l_length, attn, ids_slice, x_mask, y_mask, (z, z_p, m_p, logs_p, m_q, logs_q), (x, logw, logw_)
def infer(self, x, x_lengths, sid, tone, language, bert, noise_scale=.667, length_scale=1, noise_scale_w=0.8, max_len=None, sdp_ratio=0,y=None):
x, m_p, logs_p, x_mask = self.enc_p(x, x_lengths, tone, language, bert)
#x, m_p, logs_p, x_mask = self.enc_p(x, x_lengths, tone, language, bert)
# g = self.gst(y)
if self.n_speakers > 0:
g = self.emb_g(sid).unsqueeze(-1) # [b, h, 1]
else:
g = self.ref_enc(y.transpose(1,2)).unsqueeze(-1)
x, m_p, logs_p, x_mask = self.enc_p(x, x_lengths, tone, language, bert,g=g)
logw = self.sdp(x, x_mask, g=g, reverse=True, noise_scale=noise_scale_w) * (sdp_ratio) + self.dp(x, x_mask, g=g) * (1 - sdp_ratio)
w = torch.exp(logw) * x_mask * length_scale
w_ceil = torch.ceil(w)