deepgeek/CosyVoice
2
0
Fork 0
mirror of https://github.com/FunAudioLLM/CosyVoice.git synced 2026-02-04 17:39:25 +08:00
Code Issues Packages Projects Releases Wiki Activity

fix cross_lingual bug

Browse Source
This commit is contained in:
lyuxiang.lx
2025-04-19 09:08:47 +08:00
parent 8d67d17f73
commit e97cd1b655
2 changed files with 4 additions and 4 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
cosyvoice/cli/cosyvoice.py
View File

@@ -102,9 +102,9 @@ class CosyVoice:
yield model_output
start_time = time.time()
def inference_cross_lingual(self, tts_text, prompt_speech_16k, stream=False, speed=1.0, text_frontend=True):
def inference_cross_lingual(self, tts_text, prompt_speech_16k, zero_shot_spk_id='', stream=False, speed=1.0, text_frontend=True):
for i in tqdm(self.frontend.text_normalize(tts_text, split=True, text_frontend=text_frontend)):
model_input = self.frontend.frontend_cross_lingual(i, prompt_speech_16k, self.sample_rate)
model_input = self.frontend.frontend_cross_lingual(i, prompt_speech_16k, self.sample_rate, zero_shot_spk_id)
start_time = time.time()
logging.info('synthesis text {}'.format(i))
for model_output in self.model.tts(**model_input, stream=stream, speed=speed):

4
cosyvoice/cli/frontend.py
View File

@@ -178,8 +178,8 @@ class CosyVoiceFrontEnd:
model_input['text_len'] = tts_text_token_len
return model_input
def frontend_cross_lingual(self, tts_text, prompt_speech_16k, resample_rate):
model_input = self.frontend_zero_shot(tts_text, '', prompt_speech_16k, resample_rate)
def frontend_cross_lingual(self, tts_text, prompt_speech_16k, resample_rate, zero_shot_spk_id):
model_input = self.frontend_zero_shot(tts_text, '', prompt_speech_16k, resample_rate, zero_shot_spk_id)
# in cross lingual mode, we remove prompt in llm
del model_input['prompt_text']
del model_input['prompt_text_len']