modelscope / FunASR

A Fundamental End-to-End Speech Recognition Toolkit and Open Source SOTA Pretrained Models, Supporting Speech Recognition, Voice Activity Detection, Text Post-processing etc.
https://www.funasr.com
Other
6.99k stars 744 forks source link

cpu离线版本,客户端建立websocket请求,串行推送多个48k的语音wav,语音转写服务会重启,而且概率很高,log.txt也没见具体报错原因。生成的dmp(每次5g左右)文件会导致docker容器不断的变大。 #2204

Open janchou92 opened 1 week ago

janchou92 commented 1 week ago

Notice: In order to resolve issues more efficiently, please raise issue following the template. (注意:为了更加高效率解决您遇到的问题,请按照模板提问,补充细节)

🐛 Bug

cpu离线版本funasr-runtime-sdk-cpu-0.4.6 使用java客户端建立websocket请求 如果是推送一条48k语音文件,funasr返回后再推送一条,funasr运行会比较稳定 如果是串行推送多条48k语音文件(2条或者多条),语音转写服务会崩溃,崩溃后会自动重启,概率很高,log.txt也没见具体报错原因 生成的dmp(每次5g左右)文件会导致docker容器不断的变大。

To Reproduce

Steps to reproduce the behavior (always include the command you ran):

  1. Run cmd '....'
  2. See error

Code sample

Expected behavior

Environment

Additional context

NBStarry commented 4 days ago

怀疑和cpu占用不会释放有关,内存越占越多

goddamnVincent commented 4 hours ago

try clear cache