作者啊，你的插件为什么一更新就挂了，

[cao1101]

D:\a_work\1\s\onnxruntime\python\onnxruntime_pybind_state.cc:743 onnxruntime::python::CreateExecutionProviderInstance CUDA_PATH is set but CUDA wasn't able to be loaded. Please install the correct version of CUDA and cuDNN as mentioned in the GPU requirements page (https://onnxruntime.ai/docs/execution-providers/CUDA-ExecutionProvider.html#requirements), make sure they're in the PATH, and that your GPU is supported.，出现C盘问题，我连C盘都没有

[cao1101]

Why do I get an error every day even if I don’t have the C drive? I need to fix it once I turn on the computer, then fix it once I turn on the computer, ahhhhhhhhhhh, if I continue to search for i, my old lady will have to walk behind me.

[cao1101]

D:\a_work\1\s\onnxruntime\python\onnxruntime_pybind_state.cc:743 onnxruntime::python::CreateExecutionProviderInstance CUDA_PATH is set but CUDA wasn't able to be loaded. Please install the correct version of CUDA and cuDNN as mentioned in the GPU requirements page (https://onnxruntime.ai/docs/execution-providers/CUDA-ExecutionProvider.html#requirements), make sure they're in the PATH, and that your GPU is supported.，出现C盘问题，我连C盘都没有

是D盘

[Nourepide]

So I understand correctly that you can't run Allor because of this error, or it just appears every time you run it?

[cao1101]

So every time I repair it, it can be used normally. The repair method is to uninstall onnxruntime-gpu, and then uninstall onnxruntime. After the case, it can be used normally after installing onnxruntime. However, every time I upgrade my comfyui or restart it, the error will occur. The above error will pop up. This plug-in calls onnxruntime. So, is this error easy to fix?

[Nourepide]

It's not the first time I've gotten information that onnx is crooked on Windows systems, which is very strange since the developer of onnx is Microsoft.

I'm going to put ComfyUI on Windows today and see how it works there.

I also need more information: Windows version, video card and whether you have onnxruntime-gpu by default in requirements.txt or not

[Nourepide]

Also please check if ImageSegmentatoin node works at all when there is an error or if it doesn't affect the work.

[cao1101]

My Windows version is 11 64-bit, and my graphics card is the NVIDIA 3080 16GB version. My requirements.txt file describes the dependencies as follows: rembg, onnx.

If you want, you can install onnxruntime with GPU support onnxruntime[gpu] For Nvidia (only for GPU onnxruntime) tensorrt cuda-python When there is a problem, the ImageSegmentation node completely fails to work, showing an error related to my D drive. However, my computer only has a C drive and no D drive. Other nodes work normally.

[cao1101]

The version on my computer should be about 15 days old. I replaced the latest one because of the issue we are discussing.

[cao1101]

Another important piece of information: my CUDA version is 12.3.

[Layer-norm]

@Nourepide Windows version is not the point. The environment installed with pytorch for CUDA 12 might cause this error. They have to complied onnxruntime-gpu manually to solve this promblem when they want to use ComfyUI_windows_portable_nvidia_cu121_or_cpu.7z or other version with cu121 environment.

[Nourepide]

Can't reproduce after a few days of use with Windows - ComfyUI Portable.

Try reinstalling Allor using these instructions.

https://nourepide.github.io/ComfyUI-Allor-Doc/installation-allor.html

[cao1101]

Received. After testing, it seems that I did not install it in the correct way. Installing it in the way you recommended has solved the problem. The plug-in is very easy to use. Thank you.

[Nourepide]

If you encounter any further problems, feel free to reach out again.