Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

No module named 'upfirdn2d_plugin' #35

Closed
lz1998 opened this issue Feb 14, 2021 · 6 comments
Closed

No module named 'upfirdn2d_plugin' #35

lz1998 opened this issue Feb 14, 2021 · 6 comments

Comments

@lz1998
Copy link

lz1998 commented Feb 14, 2021

Describe the bug
A clear and concise description of what the bug is.

No module named 'upfirdn2d_plugin'
  warnings.warn('Failed to build CUDA kernels for upfirdn2d. Falling back to slow reference implementation. Details:\n\n' + str(sys.exc_info()[1]))
Setting up PyTorch plugin "upfirdn2d_plugin"... Failed!
.../stylegan2-ada-pytorch/torch_utils/ops/upfirdn2d.py:34: UserWarning: Failed to build CUDA kernels for upfirdn2d. Falling back to slow reference implementation. Details:

To Reproduce
Steps to reproduce the behavior:

python3 train.py --outdir=train_result --data=../data/dataset1.zip --gpus=8

Please copy&paste text instead of screenshots for better searchability.

Desktop (please complete the following information):

  • CUDA 11.0
  • Debian
  • GPU V100
  • torch python3 train.py --outdir=train_result --data=../data/dataset1.zip --gpus=8

Additional context
Add any other context about the problem here.

@gokhanbaydar
Copy link

I had the same problem, installed Windows 10 SDK and now its working fine.

@OvidiuSabau
Copy link

I had the same problem, installed Windows 10 SDK and now its working fine.

What do you think the issue relates to? I'm encountering it too on Linux.

@nurpax
Copy link
Contributor

nurpax commented Feb 15, 2021

Please see my comment #2 (comment)

@OvidiuSabau
Copy link

Thanks a lot for the reply! My issue ended up being that the default gcc version was 7.3, which was incompatible with cuda. Installing gcc 5 through anaconda solved this for me. Just mentioning this in case someone else encounters this too.

@Karma1628
Copy link

Thanks a lot for the reply! My issue ended up being that the default gcc version was 7.3, which was incompatible with cuda. Installing gcc 5 through anaconda solved this for me. Just mentioning this in case someone else encounters this too.

Thanks! very important! cuda version must consistent with gcc and g++ version

@mha142
Copy link

mha142 commented Apr 24, 2024

My gcc version is : 9.5.0
I solved this issue by installing the following:

conda install ninja

conda install cudatoolkit=11.1 -c conda-forge

pip install torch==1.9.1+cu111 torchvision==0.10.1+cu111 torchaudio==0.9.1 -f https://download.pytorch.org/whl/torch_stable.html

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

6 participants