How to choose the CUDA 12 version？ Can it support CUDA 12.2？

SYSTRAN / faster-whisper

Faster Whisper transcription with CTranslate2

MIT License

10k stars 841 forks source link

How to choose the CUDA 12 version？ Can it support CUDA 12.2？ #879

Open jacobtang opened 2 weeks ago

jacobtang commented 2 weeks ago

Latest versions of ctranslate2 support CUDA 12 only.

model = faster_whisper.WhisperModel(

File "/usr/local/lib/python3.10/dist-packages/faster_whisper/transcribe.py", line 144, in init self.model = ctranslate2.models.Whisper( RuntimeError: CUDA failed with error the operation cannot be performed in the present state

I have tried many times,but still have the RuntimeError. Thanks!

dancinkid6 commented 2 weeks ago

i just set it up on CUDA 12.4 and CUDNN 9.2, works perfectly, maybe u can upgrade? btw, i noticed your tesla t4 is OFF