Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

识别完字幕后闪退 #430

Closed
bailitingqin opened this issue Jun 9, 2024 · 7 comments
Closed

识别完字幕后闪退 #430

bailitingqin opened this issue Jun 9, 2024 · 7 comments

Comments

@bailitingqin
Copy link

复现方式:
1.版本:1.89(含patch)
2.语言:日译中
3.模型尝试:large-v3,tiny,small,medium都尝试了
4.环境:4090 laptop
5.参数设定:为防止是显存问题,temperature已经设成0,beam_size=1,best_of=1,condition_on_previous_text=false。
6.其它:设置cuda加速。

视频文件为2小时,最初用largev3时确实是显存问题,但是采用分割/使用小模型/改参数等方法之后,显存都用不满,但无一例外的都是识别完字幕后闪退。

cuda已经正确配置。
Cuda compilation tools, release 12.5, V12.5.40

cudnn等dll也已经copy到system32,后来尝试复现的时候还copy了一份到exe目录。

补充:
关闭cuda加速后,能够完成视频翻译。

@jianchang512
Copy link
Owner

打开 videotrans/set.ini ,搜索 temperature=1 改为 temperature=0 重启软件再试

@bailitingqin
Copy link
Author

bailitingqin commented Jun 9, 2024

额……可能是我字写的太多了吧,复现的第五点已经写了呀,temperature已经设定成0了。

@jianchang512
Copy link
Owner

这是 faster-whisper存在的问题,在执行完卸载时可能会闪退,

参考 SYSTRAN/faster-whisper#71

@jianchang512
Copy link
Owner

或者源码部署,使用整体识别模式

打开 videotrans/recognition/all.py 搜索 del model,然后将该代码全部删掉,再试试

@bailitingqin
Copy link
Author

哦……那感觉只能源码部署试试了。

python --version
Python 3.12.3

这个python版本没问题吧?应该是兼容3.10的吧。

或者我可以尝试一下打断自动化步骤,音视频转字幕->翻译字幕->字幕嵌入视频,试试。

@jianchang512
Copy link
Owner

兼容3.10

@bailitingqin
Copy link
Author

好的,谢谢

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants