search

C0untFloyd / roop-unleashed

Evolved Fork of roop with Web Server and lots of additions
GNU Affero General Public License v3.0
1.89k stars 448 forks source link

Unable to use GPU as provider for swap #619

Closed khanajan closed 4 months ago

khanajan commented 4 months ago

Describe the bug I am able to work the software while i set the provider to " CPU " and tick " Force CPU face analyzer ", but of course the conversion takes up time. I encounter issues when i set provider to "GPU", i start encountering issues as follows,

  1. Source file shows the file name but i am unable to see the image in the INPUT FACE - image of software+backend attached
  2. The preview image does not mask the face even after ticking FACE SWAP FRAMES
  3. When i START processing it works but end product is the same video without any swap.

find model: C:\roop-unleashed-main\installer\roop-unleashed\models\buffalo_l\w600k_r50.onnx recognition ['None', 3, 112, 112] 127.5 127.5 set det-size: (640, 640) 2024-04-21 11:20:08.1587507 [E:onnxruntime:, sequential_executor.cc:514 onnxruntime::ExecuteKernel] Non-zero status code returned while running Resize node. Name:'Resize_124' Status Message: CUDA error cudaErrorNoKernelImageForDevice:no kernel image is available for execution on the device

Currently i am using precision 5520, 32 RAM, contains quadro M1200

To Reproduce Steps to reproduce the behavior:

  1. Go to '...'
  2. Click on '....'
  3. Scroll down to '....'
  4. See error

Details What OS are you using?

Are you using a GPU?

Which version of roop unleashed are you using? 3.6.7

Screenshots Image not uploading image not upload backend CUDA

C0untFloyd commented 4 months ago

Your Quadro M1200 doesn't provide the functionality CUDA needs to use here. It's just not up to the task. You could try using directml as a provider but I don't think it will run much faster than using cpu.