ModuleNotFoundError: No module named 'gckn.graphs.graphs_fast'

[jjwwno]

报错ModuleNotFoundError: No module named 'gckn.graphs.graphs_fast'，已经执行了source s

[claying]

Hi @jjwwno Did you run make to install our package, following the instructions in the installation section?

[jjwwno]

I downloaded all the packages according to the md file, the problem is that executing “python setup.py build_ext --inplace” in the /gckn/graphs/ directory reports the error setup(**configuration(top_path='').todict()) TypeError: expected str, bytes or os.PathLike object, not list, can you provide specific versions of all packages?

[claying]

Sorry that this package is not maintained anymore, so it's possible that it doesn't work with the latest versions. Could you try the following versions (In particular, I believe the Numpy version >=2.0 might have broken the installation):

python=3.7
numpy=1.19
pytorch=1.9.0
pyg=2.0.2

[jjwwno]

I tried the package you provided,but I'm still encountering the same issue.Could you export the previous virtual environment into arequirements.txtfile?It would be helpful if you could also provide the server and CUDA version.I am genuinely interested in your work,and I have tried many times without success in getting it to run.

[claying]

This work was done 4 years ago with the latest compatibility updates 2 years ago during my position at my previous institute. And unfortunately, I don't have the original environment anymore. If you really want to use the package, you could try installing all the dependencies with the version released at least 2 years ago. I'm not planning to make an update in a short time.