sandrohanea / whisper.net

Whisper.net. Speech to text made simple using Whisper Models
MIT License
547 stars 84 forks source link

GPU support Error #83

Closed HillCat closed 1 year ago

HillCat commented 1 year ago

CUDA error 35 at Z:\Projects\sandro\whisper.net\whisper.cpp\ggml-cuda.cu:420: CUDA driver version is insufficient for CUDA runtime version,

In my computer there is no Z driver and path; And I installed CUDA driver is the newest from official site. I was installed whisper.net.Runtime.Cublas, Whisper.net. version 1.4.5

image

image

sandrohanea commented 1 year ago

You don't need that virtual drive Z (that is just some generic stacktrace which shows my build VM path).

Can you, please, call nvcc --version to see which version of nvidia driver you have?

Maybe try to re-install CUDA 12.1.1 (that's the latest version, used to build Whisper.net.Runtime.Cublas).

HillCat commented 1 year ago

Yes,I solved this, Downgrade the Cuda Version from 12.x to 11.x is OK, because my Nvidia card Driver is lower than 527.41. but use GPU can not increase the process speed ,why ? a 15 minute video process 10 minute use CPU, when use whisper.net.Runtime.Cublas ,there is no obvious difference.There are no advantages to using GPU.Are there any other parameters to force USE the GPU. GPU is Nvidia Gforce GT730 2GB , CPU is 11Gen i7 -11700.

image

sandrohanea commented 1 year ago

Hello @HillCat , No, only part of the transcoding process is done in GPU, same as in whisper.cpp (encoder part) => there is a differece but there will be still quite some CPU usage for now.