目录

CosyVoice

CosyVoice windows版本：

conda activate py311

pip install hyperpyyaml pip install modelscope pip install openai-whisper pip install inflect pip install ffmpeg-python

conda install pynini=2.1.5 -c conda-forge -y

测试pynini ok；

python -c "import pynini; print('pynini 版本:', pynini.__version__); print('导入成功！')"

pip install WeTextProcessing==1.0.3

下载模型：

from modelscope import snapshot_download # 下载基础模型（与您问题中名称最接近的版本） snapshot_download('iic/CosyVoice-300M', local_dir='pretrained_models/CosyVoice-300M')