Empty srt files

[phyfutima]

On Windows 10, using Purfview's Faster-Whisper model large-v3, I generated srt files from seven mp4 clips, and all seven of the resultant srt files are completely empty (and when I open the mp4 clip, it gives an error saying the associated srt file "might not be a valid subtitle file"). Why is the audio-to-text not working, am I doing something wrong?

[niksedk]

Try attaching the "whisper_log.txt" file.

Also, "large-v2" is recommended over "large-v3".

[phyfutima]

I just tried:

C:\Users\rober>whisper-faster.exe "C:\Users\rober\Documents\Ranko\060124:\Clip (June 1 2024 at 1233 PM).mp4" --language English --model medium --output_dir source

And the response was:

'whisper-faster.exe' is not recognized as an internal or external command,
operable program or batch file.

Do you know what I'm doing wrong?

[Purfview]

You are trying to run it in C:\Users\rober folder, I doubt that exe is there.

Navigate to the right folder with cd command or write full path to the exe.

[phyfutima]

Thank you, it looks like we might be getting closer, but where do I find cudnn_ops_infer64_8.dll and make sure it's in my library path?

C:\Users\rober\AppData\Roaming\Subtitle Edit\Whisper\Purfview-Whisper-Faster>whisper-faster.exe "C:\Users\rober\Documents\Ranko\060124\Clip (June 1 2024 at 1233 PM).mp4" --language English --model medium --output_dir source

Standalone Faster-Whisper r192.3 running on: CUDA

Starting work on: C:\Users\rober\Documents\Ranko\060124\Clip (June 1 2024 at 1233 PM).mp4

Could not load library cudnn_ops_infer64_8.dll. Error code 126
Please make sure cudnn_ops_infer64_8.dll is in your library path!

[Purfview]

In Subtitle Edit there should've been a popup to download those required cuBLAS and cuDNN 8.x for CUDA v11.x libs, didn't you see anything like that?
You can download them manually from there: https://github.com/Purfview/whisper-standalone-win/releases/tag/libs

@niksedk did you changed something about it, because the user bellow didn't had libs too?

[phyfutima]

That did it! Thank you so much!

[Whispered123]

Thanks.

[Purfview]

Try tiny model.

[phyfutima]

Thanks for the suggestion! I just tried that (both with tiny.en and tiny), and unfortunately same result: empty srt files. How can I troubleshoot this further, do you know?

[Whispered123]

I have the same problem with short video. But longer video, it shows "no text found".


whisper_log.txt

Model not found ? Why ?

[Whispered123]

Thx for your reply.
So can I download them manually ?

[Purfview]

So can I download them manually ?

Yes -> https://huggingface.co/Systran

[niksedk]

Should look something like this:

[phyfutima]

My Purfview-Whisper-Faster folder only has the whisper-faster app and _models folder in it (no cache, _xxl_data, faster-whisper-xxl, or ffmpeg as pictured in screenshots), could that be the problem? My models folders, such as faster-whisper-tiny.en, do have the config, model, tokenizer, and vocabulary files like yours though.

[Purfview]

could that be the problem?

No.