search

jianchang512 / ChatTTS-ui

一个简单的本地网页界面，使用ChatTTS将文字合成为语音，同时支持对外提供API接口。A simple native web interface that uses ChatTTS to synthesize text into speech, along with support for external API interfaces.
https://pyvideotrans.com
Other
5.1k stars 556 forks source link

deploy: add cuda docker-compose.yaml #41

Closed QIN2DIM closed 1 month ago

QIN2DIM commented 1 month ago

fix: #38, #3

QIN2DIM commented 1 month ago

@jianchang512 可以帮忙看下吗，运行后遇到报错

这个依赖似乎刚被移除。

chat-tts-ui exited with code 1
chat-tts-ui  | Traceback (most recent call last):
chat-tts-ui  |   File "/app/app.py", line 42, in <module>
chat-tts-ui  |     from waitress import serve
chat-tts-ui  | ModuleNotFoundError: No module named 'waitress'
QIN2DIM commented 1 month ago 
chat-tts-ui  | Traceback (most recent call last):
chat-tts-ui  |   File "/app/app.py", line 52, in <module>
chat-tts-ui  |     CHATTTS_DIR = snapshot_download('pzc163/chatTTS',cache_dir=MODEL_DIR)
chat-tts-ui  |   File "/home/venv/lib/python3.9/site-packages/modelscope/hub/snapshot_download.py", line 72, in snapshot_download
chat-tts-ui  |     os.makedirs(temporary_cache_dir, exist_ok=True)
chat-tts-ui  |   File "/usr/lib/python3.9/os.py", line 225, in makedirs
chat-tts-ui  |     mkdir(name, mode)
chat-tts-ui  | PermissionError: [Errno 13] Permission denied: '/app/models/temp'
QIN2DIM commented 1 month ago

image

QIN2DIM commented 1 month ago

@jianchang512 ok了，可以帮忙看下，我这边笔电和显卡服务器都可以正常跑。这个镜像预装了 torch-cuda 三件套和 toolkit，在构建时补装项目 requirements，正常启动即可。

jianchang512 commented 1 month ago

可以考虑修改下 README.md 添加 docker 部署方法

QIN2DIM commented 1 month ago

@jianchang512 你这边有测试过吗，我这边可以正常启，但还是有一些问题，modelscope 拉模型的时候会默认在 WORKDIR/models/temp 创建目录，这一步实际上会出现权限问题，https://github.com/jianchang512/ChatTTS-ui/pull/41#issuecomment-2143544011 可以考虑微调一下代码，或者始终以 root 运行？

我稍后提个 PR 改 README.md

jianchang512 commented 1 month ago

docker时root权限运行吧。或者提前创建temp目录