jhj0517 / Whisper-WebUI

A Web UI for easy subtitle using whisper model.
Apache License 2.0
1.2k stars 173 forks source link

Error in the middle of transcribing #281

Closed DAWiguna closed 1 month ago

DAWiguna commented 1 month ago

I use Windows 11, RTX 3080 10GB VRAM, 32GB RAM, and implement Whisper using Pinokio At first there are problem with cudnn, but after get them from cuda zip, simply unzip cudnn_ to CUDA\bin in Program Files. Then, it could run if short

Then I put our video to be subtitled. It ran for about 1 minute, then Error. But in Terminal (pinokio) there were no error messages. Whisper simply just stop in the middle.

I use model Large-v3, set Beam to 5. Any other setting that I should know about?

jhj0517 commented 1 month ago

As far as I know, Pinokio uses an older version of this WebUI and it may be problematic because of using an older version of CUDA etc.

I raised the issue to update the WebUI to the repository that pinokio may be using.

I recommend using Whisper-WebUI-Portable-Windows.zip for now.

DAWiguna commented 1 month ago

Thanks, I guess that's the best.