My IDE is vs code, and my pytorch version is 0.4 without GPU.
When I import other extension package like numpy, my vs code setting is, when my mouse move to package’s functions, vs code will pop up function’s docstring to let me get understand the API.
However, when I import torch, pylint pop up a error, which is: missing module docstring.
And all functions in torch cannot pop up explanation. When I use “Ctrl+clip” to find the function, it seems vs code cannot find out. But these torch’s functions still work.
I don’t know is it torch has not docstrings or just my version’s problem.
I can’t put up the image, so I hope you can understand my descriptions.
The problem is known and will be fixed in next release as you can see here
Thanks a lot! But I think the problem is kind a of different. Although I met the same problem proposed in the link, and I set the pylint to ignore these inspections.
My problem is functions in pytorch such as torch.randn, has no interpretation, and I can’t jump to the source code of torch.randn as well.
I just want to make sure that is it there are some mistakes in my configuration or is it pytorch has not such functional interpretation in source code?
That should be a part of this problem. I don’t know what exactly you are looking for but code inspections with IDEs should be (hopefully) completely fixed in next release
I think this is part of the missing autocomplete feature here, but I am not really sure. Maybe @colesbury can verify this?