Medix-AI/coqui-tts
Medix-AI
/
coqui-tts
mirror of https://github.com/coqui-ai/TTS.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

Disable the reversal prosody encoder speaker loss

This commit is contained in:
Edresson Casanova 2022-05-19 13:55:14 +00:00
parent 7b85703b28
commit 63d27bc8d4
1 changed files with 9 additions and 6 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

15
TTS/tts/models/vits.py
View File

@ -509,6 +509,7 @@ class VitsArgs(Coqpit):
prosody_embedding_dim: int = 0
prosody_encoder_num_heads: int = 1
prosody_encoder_num_tokens: int = 5
use_prosody_enc_spk_reversal_classifier: bool = False
detach_dp_input: bool = True
use_language_embedding: bool = False
@ -647,11 +648,12 @@ class Vits(BaseTTS):
num_style_tokens=self.args.prosody_encoder_num_tokens,
gst_embedding_dim=self.args.prosody_embedding_dim,
)
self.speaker_pros_enc_reversal_classifier = ReversalClassifier(
in_channels=self.args.prosody_embedding_dim,
out_channels=self.num_speakers,
hidden_channels=256,
)
if self.args.use_prosody_enc_spk_reversal_classifier:
self.speaker_reversal_classifier = ReversalClassifier(
in_channels=self.args.prosody_embedding_dim,
out_channels=self.num_speakers,
hidden_channels=256,
)
if self.args.use_text_enc_spk_reversal_classifier:
self.speaker_text_enc_reversal_classifier = ReversalClassifier(
@ -996,7 +998,8 @@ class Vits(BaseTTS):
l_pros_speaker = None
if self.args.use_prosody_encoder:
pros_emb = self.prosody_encoder(z).transpose(1, 2)
_, l_pros_speaker = self.speaker_pros_enc_reversal_classifier(pros_emb.transpose(1, 2), sid, x_mask=None)
if self.args.use_prosody_enc_spk_reversal_classifier:
_, l_pros_speaker = self.speaker_reversal_classifier(pros_emb.transpose(1, 2), sid, x_mask=None)
x, m_p, logs_p, x_mask = self.text_encoder(x, x_lengths, lang_emb=lang_emb, emo_emb=eg, pros_emb=pros_emb)