deepgeek/CosyVoice
2
0
Fork 0
mirror of https://github.com/FunAudioLLM/CosyVoice.git synced 2026-02-05 18:09:24 +08:00
Code Issues Packages Projects Releases Wiki Activity

add llm bistream

Browse Source
This commit is contained in:
lyuxiang.lx
2025-01-23 10:12:06 +08:00
parent 0b75c3a03f
commit 07e477519b
10 changed files with 163 additions and 39 deletions
Download Patch File Download Diff File Expand all files Collapse all files

3
cosyvoice/cli/cosyvoice.py
View File

@@ -13,6 +13,7 @@
# limitations under the License.
import os
import time
from typing import Generator
from tqdm import tqdm
from hyperpyyaml import load_hyperpyyaml
from modelscope import snapshot_download
@@ -76,7 +77,7 @@ class CosyVoice:
def inference_zero_shot(self, tts_text, prompt_text, prompt_speech_16k, stream=False, speed=1.0, text_frontend=True):
prompt_text = self.frontend.text_normalize(prompt_text, split=False, text_frontend=text_frontend)
for i in tqdm(self.frontend.text_normalize(tts_text, split=True, text_frontend=text_frontend)):
if len(i) < 0.5 * len(prompt_text):
if (not isinstance(i, Generator)) and len(i) < 0.5 * len(prompt_text):
logging.warning('synthesis text {} too short than prompt text {}, this may lead to bad performance'.format(i, prompt_text))
model_input = self.frontend.frontend_zero_shot(i, prompt_text, prompt_speech_16k, self.sample_rate)
start_time = time.time()