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

sum style tokesn with encoder outputs instead of concat

This commit is contained in:
Eren Golge 2019-09-20 18:46:59 +02:00
parent 14a4d1a061
commit 6561013d28
1 changed files with 6 additions and 6 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

12
models/tacotrongst.py
View File

@ -56,14 +56,14 @@ class TacotronGST(Tacotron):
self._init_states()
self.compute_speaker_embedding(speaker_ids)
if self.num_speakers > 1:
inputs = self._concat_speaker_embedding(inputs,
inputs = self._add_speaker_embedding(inputs,
self.speaker_embeddings)
encoder_outputs = self.encoder(inputs)
if self.num_speakers > 1:
encoder_outputs = self._concat_speaker_embedding(encoder_outputs,
encoder_outputs = self._add_speaker_embedding(encoder_outputs,
self.speaker_embeddings)
gst_outputs = self.gst(mel_specs)
encoder_outputs = self._concat_speaker_embedding(
encoder_outputs = self._add_speaker_embedding(
encoder_outputs, gst_outputs)
mel_outputs, alignments, stop_tokens = self.decoder(
encoder_outputs, mel_specs, mask, self.speaker_embeddings_projected)
@ -78,16 +78,16 @@ class TacotronGST(Tacotron):
self._init_states()
self.compute_speaker_embedding(speaker_ids)
if self.num_speakers > 1:
inputs = self._concat_speaker_embedding(inputs,
inputs = self._add_speaker_embedding(inputs,
self.speaker_embeddings)
encoder_outputs = self.encoder(inputs)
if self.num_speakers > 1:
encoder_outputs = self._concat_speaker_embedding(encoder_outputs,
encoder_outputs = self._add_speaker_embedding(encoder_outputs,
self.speaker_embeddings)
if style_mel is not None:
gst_outputs = self.gst(style_mel)
gst_outputs = gst_outputs.expand(-1, encoder_outputs.size(1), -1)
encoder_outputs = self._concat_speaker_embedding(encoder_outputs,
encoder_outputs = self._add_speaker_embedding(encoder_outputs,
gst_outputs)
mel_outputs, alignments, stop_tokens = self.decoder.inference(
encoder_outputs, self.speaker_embeddings_projected)