Help, the software is not working! about whisper-ctranslate2

[jordimas]

Which version of whisper-ctranslate2 are you using?
Are you using CPU or GPU?

[boykan313]

Какую версию шепота-ctranslate2 вы используете? Вы используете процессор или графический процессор?

I am using the latest version of your software (downloaded from the header, used the install command via pip.
) and I am using a GPU. cuda and the rest of the software is installed, except for NVIDIA cuBLAS, at the moment it does not have support on windows, but this does not prevent me from using gpu on a regular visper, which confirms my screen.

[jordimas]

I do not own a Windows box. I cannot help you with Windows specific issues.

What I can say is make sure that you are running whisper-ctranslate 0.26 which was release today, since it contains a Windows fix.

[boykan313]

I do not own a Windows box. I cannot help you with Windows specific issues.

What I can say is make sure that you are running whisper-ctranslate 0.26 which was release today, since it contains a Windows fix.

does the new version need to be installed in some special way or is it enough to type the command "pip install -U whisper-ctranslate2"?

[boykan313]

I do not own a Windows box. I cannot help you with Windows specific issues.

What I can say is make sure that you are running whisper-ctranslate 0.26 which was release today, since it contains a Windows fix.

tell me how can I make sure that this version works for me, if nothing happens after piercing this command. For example, having broken through the whisper command, there is a reaction.

Thank you for your reply, I really want to understand this topic.

[jordimas]

Try

whisper-ctranslate2 --version

it should show:

whisper-ctranslate2 0.2.6

If it's older, try:

pip3 install whisper-ctranslate2==0.2.6

[boykan313]

Version 0.2.6 is installed, but still nothing happens after the main command is executed, it hangs on the language detection.

[jjj001]

I have similar problem.

I tried to process mp3 file, it downloaded model, detected language and after a short while finished without any error message nor output file. It does even with very short mp3 file and tiny database.

[koenigscode]

a uni colleague had the same problem (windows), he has an NVIDIA gpu in his laptop (surface laptop studio), and it defaulted to cuda, but same as here it just stuck at detecting the language.
explicitly setting --device cpu worked (huge performance loss i suppose, but hey, at least it runs)

used version 0.2.9

so I guess this issue is similar to #40