subprocess.CalledProcessError: Command '['ninja', '-v']'

[ghost]

hi,

when executing python train.py --evaluate --snapshot checkpoints/best_cityscapes_checkpoint.pth

I get this error: subprocess.CalledProcessError: Command '['ninja', '-v']' returned non-zero exit status 1.

I'm using: gcc=4.9.4 ninja=1.9.0 pytorch = 1.2.0 torchvision = 0.4.0 python = 3.6.9 cuda = 9.2 torch-encoding = 1.0.0

may you share the package versions of your environment?

[kaushikb258]

I am having the same issue. Did you fix it?

[tovacinni]

https://github.com/nv-tlabs/GSCNN/pull/10

The dependencies listed in the Dockerfile here should work. Can you try it?

[kaushikb258]

I got the code working by setting syncbn to False as I only have 1 GPU anyway, and evaluate works. I tried the Dockerfile, but the Docker was not building. Thanks anyway!

[ghost]

Hi,

it is still not working for me. So I applied @kaushikb258 workaround to get it working

[yuki-0321]

1.Reference：https://zhuanlan.zhihu.com/p/53418563

Edit the file : /usr/local/lib/python3.7/site-packages/torch/utils/cpp_extension.py You can try to change ['ninja', '-v'] to ['ninja', '--version']

2.You can try to install pytorch = 1.1.0. This may be a issue of pytorch's version.

I'm using:

ninja=1.9.0 torch = 1.1.0 torchvision = 0.2.0 python = 3.7.3 cuda = 10.1 torch-encoding = 1.0.0

[paul-adlink]

Dear all, I'm using python3.6 and also met this problem. I fixed this problem by modify the /usr/local/lib/python3.6/site-packages/torch/utils/cpp_extension.py from ['ninja', '-v'] to ['ninja', '--version']. Hope this information could help you who met this problem also.

And later I met no module enclib_cpu problem. I rollback to the torch version==1.1.0 like @yuki-0321 mentioned. The problem seems solved and without doing any extra modify about ninja --version.

Hope this information is helpful for who also raise this kind of problems.