CUDA device error "module 'torch._C' has no attribute '_cuda_setDevice'" when running test.py

[FoxTrotte]

Hi !

Trying out test.py results in the following error:

Traceback (most recent call last): File "test.py", line 26, in <module> torch.cuda.set_device(0) File "C:\Users\natha\anaconda3\envs\ColorVid\lib\site-packages\torch\cuda\__init__.py", line 311, in set_device torch._C._cuda_setDevice(device) AttributeError: module 'torch._C' has no attribute '_cuda_setDevice'

I tried installing pytorch manually using their tool https://pytorch.org/get-started/locally/ (with CUDA 11.6) but that doesn't resolve the issue.

Can someone help me understand what is going on ? Thanks !!

[HolyMountains]

I have the same Issue! Do you found a solution?

[xixifusi1984]

I have the same Issue! Did some body get the key to fix it?

[UpscaleHistory]

It took me three days but I found a fix for my computer:
Downgrade python to 3.6, and install pytorch 1.6.0 with:
pip install torch===1.6.0 torchvision===0.7.0 -f https://download.pytorch.org/whl/torch_stable.html

If you open the colab demo, you will see "# seems to be a colab bug, need to install previous version for pytorch".
Then the notebook proceeds to install torch 1.6.0 with pip install torch===1.6.0 torchvision===0.7.0.
It's the same bug with my local machine, except that I can't install torch 1.6.0 on my computer using the same code.
When use the same code with conda, I get "ERROR: No matching distribution found for torch==1.6.0"
I think it's because colab is using Tesla T4, where as my computer is rtx 2070.

[UpscaleHistory]

If you are trying to run the code locally with anaconda or connect to a local runtime with colab, you might need to change
pip install torch==1.6.0 torchvision==0.7.0
to
pip install torch===1.6.0 torchvision===0.7.0 -f https://download.pytorch.org/whl/torch_stable.html
depending on your gpu.
The first code works with colab's T4, but not my rtx 2070; the second code works for both.
I think colab A100 also has some kind of issues, but it's related to cuda.