search

Anjok07 / ultimatevocalremovergui

GUI for a Vocal Remover that uses Deep Neural Networks.
MIT License
16.56k stars 1.25k forks source link

Not using GPU #880

Open diegomora02 opened 9 months ago

diegomora02 commented 9 months ago

image image

I have a laptop with an Nvidia GTX 1650 and it is not using it despite having GPU processing enabled, I have downloaded the most recent drivers but it is clearly using CPU

AngusGhoul commented 8 months ago

I have the same fuckin problem and it´s really annoying, did you find the solution?

Anjok07 commented 8 months ago

@AngusGhoul I made some important changes the latest release that should have resolved this issue for most people. What type of GPU do you have?

@diegomora02 A majority of models in UVR require at least 8GB's of V-RAM (6 being the absolute bare minimum) and it looks like you only have 4. In some cases if the application can't run on the GPU, it defaults to the CPU instead.

AngusGhoul commented 8 months ago

@Anjok07 yes I start to have this problem after I updated yesterday. I have a Geforce 3050rtx. but it´s weird because works sometimes, and sometimes not.

AngusGhoul commented 8 months ago

this one is my version Ultimate Vocal Remover v5.6.0 [2023-10-18 13:58:59]

Anjok07 commented 8 months ago

@AngusGhoul Try these 2 things

  1. If you're using MDX-Net, choose any segment size other than 256
  2. Choose your GPU directly (see image). If you don't see a GPU selection that doesn't look like the image below, please reinstall the patch here choose_gpu
AngusGhoul commented 8 months ago

Yes, I selected the gpu manually before. I will send you screenshots of my task manager. Like I said, the program was working before the last patch. I dont know what to do but was working before :( imagen_2023-10-20_205203096 imagen_2023-10-20_205237900 imagen_2023-10-20_205300449

VamuveTV commented 6 months ago

Sorry, i didn´t understood this. I don´t have a graphic card that uses 8 Gb. Insetad, mine is only of 2 Gb (NVidia Geforce GT 1030 2Gb) and i´m using 48 Gb of RAM memory. What are the correct configurations for using UVR with my hardware ?

Can you update UVR to it automatically detect and configure for the user´s hardware ?

Clipboard01

ChrisTG742 commented 6 months ago

I think a 1030 has a too old revision of CUDA. But even if it would work, the CUDA-cores do not have enough power to bring any performance boost compared to a decent CPU. I believe everything below GT1080 will not make sense.

lannodev commented 5 months ago

@Anjok07 I have the same problem using Windows Manual Installation, Is there a way to apply the last patch on manual installation?