Cant run on windows

[hamaadtahiir]

When I try to run the inference it gives the following error:

Loading networks from "data/network-snapshot-004000.pkl"...
Loading custom kernel...
Traceback (most recent call last):
File "test.py", line 159, in
generate_images() # pylint: disable=no-value-for-parameter
File "g:\c-vton\pfafn\lib\site-packages\click\core.py", line 1128, in call
return self.main(*args, **kwargs)
File "g:\c-vton\pfafn\lib\site-packages\click\core.py", line 1053, in main
rv = self.invoke(ctx)
File "g:\c-vton\pfafn\lib\site-packages\click\core.py", line 1395, in invoke
return ctx.invoke(self.callback, **ctx.params)
File "g:\c-vton\pfafn\lib\site-packages\click\core.py", line 754, in invoke
return __callback(*args, **kwargs)
File "g:\c-vton\pfafn\lib\site-packages\click\decorators.py", line 26, in new_func
return f(get_current_context(), *args, **kwargs)
File "test.py", line 93, in generate_images
G = legacy.load_network_pkl(f)['G_ema'].to(device) # type: ignore
File "G:\c-vton\PASTA-GAN\legacy.py", line 21, in load_network_pkl
data = _LegacyUnpickler(f).load()
File "G:\c-vton\PASTA-GAN\torch_utils\persistence.py", line 191, in _reconstruct_persistent_obj
module = _src_to_module(meta.module_src)
File "G:\c-vton\PASTA-GAN\torch_utils\persistence.py", line 227, in _src_to_module
exec(src, module.dict) # pylint: disable=exec-used
File "", line 2243, in
NameError: name 'os' is not defined

Now os module is obviously installed and importable in python

[Deleted user]

Ensure you have properly followed the installation instructions. Make sure you are running the script from inside the virtual environment, and that you have the specified version of torch installed.

[miibotree]

I think the problem is the cuda version on your environment is not equal to the requirement 11.0.
See Line1204-1225 on networks.py:

from torch.autograd import Function
from torch.utils.cpp_extension import load

def is_custom_kernel_supported():
    version_str = str(torch.version.cuda).split(".")
    major = version_str[0]
    minor = version_str[1]
    return int(major) >= 10 and int(minor) >= 1

if is_custom_kernel_supported():
    print("Loading custom kernel...")
    module_path = os.path.dirname(__file__)
    upfirdn2d_op = load(
        'upfirdn2d',
        sources=[
            os.path.join(module_path, 'upfirdn2d.cpp'),
            os.path.join(module_path, 'upfirdn2d_kernel.cu'),
        ],
        verbose=True
    )

use_custom_kernel = is_custom_kernel_supported()

if your cuda version satisfied with the condition : int(major) >= 10 and int(minor) >= 1, will go to the above codes.
However, cuda version 11.0 not satisfied with the condition and thus not go to the above codes.
By the way, @alexlevenston2021 I didn't find file upfirdn2d_kernel.cu in the source code? Do I miss compile any thing?

[omking]

@miibotree please rename PASTA-GAN-main/torch_utils/ops/upfirdn2d.cu to upfirdn2d_kernel.cu it will solve your issue

after solving this issue I stuck in
RuntimeError: Error building extension 'upfirdn2d'

any able to run full code for test?