0.环境配置# 创建conda环境 conda create -n moss-tts-nano python3.12 -y conda activate moss-tts-nano # 下载源码 git clone https://github.com/OpenMOSS/MOSS-TTS-Nano.git cd MOSS-TTS-Nano # 安装依赖 pip install -r requirements.txt # 纯本地跑源码不改包结构,可以不用 pip install -e . pip install -e .1.版本问题解决ERROR: pips dependency resolver does not currently take into account all the packages that are installed. This behaviour is the source of the following dependency conflicts. soundfile 0.14.0 requires cffi1.0, which is not installed.soundfile 依赖 cffi1.0 没装直接装上就行pipinstallcffiRuntimeError: No module named importlib_resourcespipinstallimportlib-resources2.连接不上huggingface问题解决Max retries exceeded with url: /OpenMOSS-Team/MOSS-TTS-Nano/resolve/main/config.json (Caused by ConnectTimeoutError(HTTPSConnection(hosthuggingface.co, port443) at 0x1f5f5c7ffb0, Connection to huggingface.co timed out. (connect timeout10)))),临时解决当前终端生效WindowsCMDset HF_ENDPOINThttps://hf-mirror.com python infer.py --prompt-audio-path assets/audio/zh_1.wav --text 欢 迎关注模思智能、上海创智学院与复旦大学自然语言处理实验室。WindowsPowerShell$env:HF_ENDPOINT https://hf-mirror.compython infer.py--prompt-audio-path assets/audio/zh_1.wav--text 欢 迎关注模思智能、上海创智学院与复旦大学自然永久解决以后不用每次输Windows 系统环境变量 搜索 “环境变量” 系统变量 → 新建 变量名HF_ENDPOINT 变量值https://hf-mirror.com 确定 → 重启终端 / IDE如果模型已经缓存到本地不想再联网set HF_HUB_OFFLINE13.pynini版本差 0.0.1基本不影响运行不用强行升级现在已经装好importlib-resources直接跑命令即可$env:HF_ENDPOINT https://hf-mirror.compython infer.py--prompt-audio-path assets/audio/zh_1.wav--text 欢 迎关注模思智能、上海创智学院与复旦大学自然语言
OpenMOSS / MOSS-TTS-Nano TTS文字转语音windows本地部署
发布时间：2026/6/26 4:28:11
0.环境配置# 创建conda环境 conda create -n moss-tts-nano python3.12 -y conda activate moss-tts-nano # 下载源码 git clone https://github.com/OpenMOSS/MOSS-TTS-Nano.git cd MOSS-TTS-Nano # 安装依赖 pip install -r requirements.txt # 纯本地跑源码不改包结构,可以不用 pip install -e . pip install -e .1.版本问题解决ERROR: pips dependency resolver does not currently take into account all the packages that are installed. This behaviour is the source of the following dependency conflicts. soundfile 0.14.0 requires cffi1.0, which is not installed.soundfile 依赖 cffi1.0 没装直接装上就行pipinstallcffiRuntimeError: No module named importlib_resourcespipinstallimportlib-resources2.连接不上huggingface问题解决Max retries exceeded with url: /OpenMOSS-Team/MOSS-TTS-Nano/resolve/main/config.json (Caused by ConnectTimeoutError(HTTPSConnection(hosthuggingface.co, port443) at 0x1f5f5c7ffb0, Connection to huggingface.co timed out. (connect timeout10)))),临时解决当前终端生效WindowsCMDset HF_ENDPOINThttps://hf-mirror.com python infer.py --prompt-audio-path assets/audio/zh_1.wav --text 欢 迎关注模思智能、上海创智学院与复旦大学自然语言处理实验室。WindowsPowerShell$env:HF_ENDPOINT https://hf-mirror.compython infer.py--prompt-audio-path assets/audio/zh_1.wav--text 欢 迎关注模思智能、上海创智学院与复旦大学自然永久解决以后不用每次输Windows 系统环境变量 搜索 “环境变量” 系统变量 → 新建 变量名HF_ENDPOINT 变量值https://hf-mirror.com 确定 → 重启终端 / IDE如果模型已经缓存到本地不想再联网set HF_HUB_OFFLINE13.pynini版本差 0.0.1基本不影响运行不用强行升级现在已经装好importlib-resources直接跑命令即可$env:HF_ENDPOINT https://hf-mirror.compython infer.py--prompt-audio-path assets/audio/zh_1.wav--text 欢 迎关注模思智能、上海创智学院与复旦大学自然语言
)
