tamarott / SinGAN

Official pytorch implementation of the paper: "SinGAN: Learning a Generative Model from a Single Natural Image"
https://tamarott.github.io/SinGAN.htm
Other
3.31k stars 611 forks source link

RuntimeError: Cannot initialize CUDA without ATen_cuda library. #127

Open batuhan3526 opened 4 years ago

batuhan3526 commented 4 years ago

I got this error. What should ı do ?

RuntimeError: Cannot initialize CUDA without ATen_cuda library. PyTorch splits its backend into two shared libraries: a CPU library and a CUDA library; this error has occurred because you are trying to use some CUDA functionality, but the CUDA library has not been loaded by the dynamic linker for some reason. The CUDA library MUST be loaded, EVEN IF you don't directly use any symbols from the CUDA library! One common culprit is a lack of -Wl,--no-as-needed in your link arguments; many dynamic linkers will delete dynamic library dependencies if you don't depend on any of their symbols. You can check if this has occurred by using ldd on your binary to see if there is a dependency on *_cuda.so library.

batuhan3526 commented 4 years ago

ı tried this. but didnt work

pip install torch==1.4.0+cpu torchvision==0.5.0+cpu -f https://download.pytorch.org/whl/torch_stable.html

LeviDPC commented 4 years ago

If your computer is not cuda capable ready (or if it is but you're fine not using cuda and just letting it go slow) then use the --not_cuda flag when running SinGAN. If it is cuda ready then the problem probably is with your setup not with SinGAN and I would start Googling.

Best, Clef

chhh commented 2 years ago

Same problem. I do have a CUDA card. Tried bootstrapping python with a conda environment:

name: env-singan
   dependencies:
   - python=3.6
   - pip
   - pip:
     - -r requirements.txt

then conda env create - installs python 3.6 + pip-installs all the modules. Got the ATen_cuda error.

Ezharjan commented 1 year ago

Same again...