DML device: CPU not GPU

[spoonsc2]

Hello!

I've scoured the forums so I'm unsure what I'm doing wrong.

I followed this #186 (comment) which has been reposted numerous times but when I run it gives me an error, something about Numpy version issuses, recommends downgrading to <2. The onnxruntime install upgrades Numpy to v 2.1.1 i believe.

When I run the installer, and update it does that and I get the following:

My question is, should the Device be showing as GPU not CPU? If so, help :(

Getting about 5-6 fps with 256 upscale GPEN 0.4 with a 7900xtx and 7800x3d. (currently seeing pikes sup to 65fps)

Thanks in advance.

[fahadshery]

I have the same issue. Getting the following:

docker logs -f roop-unleashed
Using provider ['CPUExecutionProvider'] - Device:cpu
/usr/local/lib/python3.10/dist-packages/gradio/components/dropdown.py:188: UserWarning: The value passed into gr.Dropdown() is not in the list of choices. Please update the list of choices to include: cuda or set allow_custom_value=True.
  warnings.warn(
Running on local URL:  http://0.0.0.0:7860

To create a public link, set `share=True` in `launch()`.

[fahadshery]

added more info here #872

[C0untFloyd]

@fahadshery You're not having the same issue as this guy. He is using DirectML where you just installed the wrong cuda version.

@spoonsc2 Yes, it will use your gpu although it's showing cpu and 5-6 fps with gpen is quite fast!

[Pr0St0ck]

IF DML NOT GPU USAGE steps:

conda activate ru

pip uninstall onnxruntime onnxruntime-directml

pip install onnxruntime-directml==1.15.1