[2noise/ChatTTS]core.py升级,插件不兼容

2024-06-05 763 views
2

python=3.10 pip install -r 正常 pip install nemo_tools 正常

代码 ‘’‘ import ChatTTS import soundfile as sf

chat = ChatTTS.Chat() chat.load_models() texts = ["你好",] wavs = chat.infer(texts, use_decoder=True) sf.write(r"D:\Desktop\chattts.wav", wavs[0][0], 24000) ’‘’

报错内容 ‘’‘ Traceback (most recent call last): File "/Users/user/Downloads/ChatTTS/import ChatTTS.py", line 7, in wavs = chat.infer(texts, use_decoder=True) File "/Users/user/Downloads/ChatTTS/ChatTTS/core.py", line 144, in infer self.init_normalizer(_lang) File "/Users/user/Downloads/ChatTTS/ChatTTS/core.py", line 181, in init_normalizer from nemo_text_processing.text_normalization.normalize import Normalizer ModuleNotFoundError: No module named 'nemo_text_processing' ’‘’

回答

4

omegaconf也有问题报错

3

同。这个包装起来好多兼容问题。查了下好像是用来解决数字/标点符号转口语问题的

1

由于依赖项Pynini在 Linux x86_64 以外的平台上,从 Pip 安装会尝试从头开始编译 Pynini

要在window上安装nemo_text_processing

先运行:conda install -c conda-forge pynini=2.1.5 再运行:pip install nemo_text_processing

6

mac还是报错,无法安装成功pynini

5
iShot_2024-05-31_22 19 54