jianchang512 / pyvideotrans

Translate the video from one language to another and add dubbing. 将视频从一种语言翻译为另一种语言,并支持api调用
https://pyvideotrans.com
GNU General Public License v3.0
10.21k stars 1.13k forks source link

使用distil-whisper-large-v3 模型 后使用标准功能后,翻译到一半后就后面这个应用程序就自动关闭了这个如何解决,已经更新到最新exe版本 #419

Closed Lucas-cen-ai closed 3 months ago

jianchang512 commented 4 months ago

查看闪退问题 https://pyvideotrans.com/11.html

Lucas-cen-ai commented 4 months ago

大佬,我看了一下log出现了以下问题 DEBUG:faster_whisper:Processing segment at 00:27.380 DEBUG:faster_whisper:Log probability threshold is not met with temperature 0.0 (-1.009665 < -1.000000) DEBUG:faster_whisper:Compression ratio threshold is not met with temperature 0.2 (2.560000 > 2.400000) DEBUG:faster_whisper:Log probability threshold is not met with temperature 0.2 (-1.225621 < -1.000000) DEBUG:faster_whisper:Log probability threshold is not met with temperature 0.4 (-1.026379 < -1.000000) DEBUG:faster_whisper:Log probability threshold is not met with temperature 0.6 (-1.083090 < -1.000000) DEBUG:faster_whisper:Log probability threshold is not met with temperature 0.8 (-1.120015 < -1.000000) DEBUG:faster_whisper:Log probability threshold is not met with temperature 1.0 (-1.687015 < -1.000000) INFO:VideoTrans:19 这个是表示什么意思呢麻烦您帮解答一下谢谢

jianchang512 commented 4 months ago

打开 videotrans/set.ini 将 temperature=1 改为 temperature=0 再试试

Lucas-cen-ai commented 4 months ago

大佬,我更改了set.ini中temperature=1 改为 temperature=0 就启动不了了

Lucas-cen-ai commented 4 months ago

log里面说的是这样的 ERROR:VideoTrans:set.ini 中有语法错误:not enough values to unpack (expected 2, got 1) ERROR:VideoTrans:set.ini 中有语法错误:not enough values to unpack (expected 2, got 1) 我写的内容如下 temperature=0,就是简单的换了应该数字,是不是我的操作有误,或者是需要什么特殊手段重启一下呢?

jianchang512 commented 4 months ago

截图你改的。

jianchang512 commented 4 months ago

还改为1吧 这个不是报错。

只是日志输出,温度不合适

jianchang512 commented 4 months ago

set.ini中如果 condition_on_previous_text 不是 false,就改为 false,是就不需要动

Lucas-cen-ai commented 4 months ago

set.ini中如果 condition_on_previous_text 不是 false,就改为 false,是就不需要动

是false的

jianchang512 commented 4 months ago

DEBUG 不是报错,可以忽略

如果GPU独显小于12G,不要使用v3模型

Lucas-cen-ai commented 4 months ago

那集显是16g的推荐使用那个语言模型会更合适呢大佬

jianchang512 commented 4 months ago

集显等于没显卡,纯CPU运行

small medium

Lucas-cen-ai commented 4 months ago

n卡有应该4g的运行内存,这种推荐那种语言模型最合适呀大佬

jianchang512 commented 4 months ago

最大也就是medium,如果视频也比较大,可能会闪退,如果闪退可以尝试 预先分割 模式