# CosyVoice **Repository Path**: lulendi/CosyVoice ## Basic Information - **Project Name**: CosyVoice - **Description**: 这是用于fork cosyvoice同步的分支,pro分支添加了生成字幕文件,新增音色和保存当前音色功能 环境使用: conda 24.11.3 python 3.11.11 NVIDIA-SMI 545.84 Driver Version: 545.84 CUDA Version: 12.3 - **Primary Language**: Python - **License**: Apache-2.0 - **Default Branch**: pro - **Homepage**: None - **GVP Project**: No ## Statistics - **Stars**: 0 - **Forks**: 1 - **Created**: 2025-03-08 - **Last Updated**: 2025-04-09 ## Categories & Tags **Categories**: Uncategorized **Tags**: None ## README 这是用于fork cosyvoice同步的分支,pro分支添加了生成字幕文件,新增音色和保存当前音色功能 环境使用: ```txt conda 24.11.3 python 3.11.11 NVIDIA-SMI 545.84 Driver Version: 545.84 CUDA Version: 12.3 ``` 安装说明: ```shell git clone --recursive https://gitee.com/lulendi/CosyVoice.git # 进入项目目录 cd CosyVoice # 进入third_party目录 cd third_party # 拉取Matcha-TTS git clone https://github.com/shivammehta25/Matcha-TTS.git # 回到项目目录 cd .. # 创建虚拟环境 建议用pycharm 右下角 No Interpreter -> Add new Interpreter 添加conda环境 conda create -n cosyvoice -y python=3.11 # 激活虚拟环境 conda activate cosyvoice # 安装pynini conda install -y -c conda-forge pynini==2.1.5 # 安装依赖 # 特别提醒 安装过程中,依赖下载 pip install -r requirements.txt -i https://mirrors.aliyun.com/pypi/simple/ --trusted-host=mirrors.aliyun.com # 下载模型 python downloadModel.py # 复制 spk2info.pt 放入pretrained_models/CosyVoice2-0.5B目录下 ``` GPU诊断报告: ```shell python gpu_diagnostics.py ``` 待完成: ```shell api接口整合 ``` 生成requirements.txt ```shell pipreqs ./ --encoding=utf8 --force ```