mirror of
https://github.com/FunAudioLLM/CosyVoice.git
synced 2026-02-04 17:39:25 +08:00
@@ -149,6 +149,11 @@ class CosyVoiceFrontEnd:
|
||||
prompt_speech_resample = torchaudio.transforms.Resample(orig_freq=16000, new_freq=resample_rate)(prompt_speech_16k)
|
||||
speech_feat, speech_feat_len = self._extract_speech_feat(prompt_speech_resample)
|
||||
speech_token, speech_token_len = self._extract_speech_token(prompt_speech_16k)
|
||||
if resample_rate == 24000:
|
||||
# cosyvoice2, force speech_feat % speech_token = 2
|
||||
token_len = min(int(speech_feat.shape[1] / 2), speech_token.shape[1])
|
||||
speech_feat, speech_feat_len[:] = speech_feat[:, :2 * token_len], 2* token_len
|
||||
speech_token, speech_token_len[:] = speech_token[:, :token_len], token_len
|
||||
embedding = self._extract_spk_embedding(prompt_speech_16k)
|
||||
model_input = {'text': tts_text_token, 'text_len': tts_text_token_len,
|
||||
'prompt_text': prompt_text_token, 'prompt_text_len': prompt_text_token_len,
|
||||
|
||||
@@ -380,8 +380,7 @@ class CosyVoice2Model:
|
||||
while True:
|
||||
time.sleep(0.1)
|
||||
if len(self.tts_speech_token_dict[this_uuid]) - token_offset >= self.token_hop_len + self.flow.pre_lookahead_len:
|
||||
this_tts_speech_token = torch.tensor(self.tts_speech_token_dict[this_uuid][:token_offset + self.token_hop_len + self.flow.pre_lookahead_len]) \
|
||||
.unsqueeze(dim=0)
|
||||
this_tts_speech_token = torch.tensor(self.tts_speech_token_dict[this_uuid][:token_offset + self.token_hop_len + self.flow.pre_lookahead_len]).unsqueeze(dim=0)
|
||||
this_tts_speech = self.token2wav(token=this_tts_speech_token,
|
||||
prompt_token=flow_prompt_speech_token,
|
||||
prompt_feat=prompt_speech_feat,
|
||||
|
||||
Reference in New Issue
Block a user