Request for comprehensive documentation for Libtorch

Dear Pytorch Team,

Pytorch’s documentation is probably one of the best I have ever seen.
However, the case with Libtorch is rather sad!!!
There are just some very basic examples which seems pretty random and not well organized.
I understand that it follows Pytorch in terms of use.

However, to really make it a solid framework for C++ usage, I would request a thorough documentation for the C++ frontend as well. At the very least - each API needs at least a 1 line example.

Presently we have to actually see the source code headers and try to use it.
This is not very nice.

I would really love to see Pytorch/Libtorch as a single solution for all deep learning needs.

Best Regards

Thanks for the feedback as I agree that the C++ docs could be improved.
Contributions in this area are more than welcome so would you be interested in improving (some of) these docs?

Yes, surely. I would love to contribute my 2 cents whenever possible.
How may I add/ edit documentation?

The libtorch docs should be here and you could add more documentation there.
I would recommend to create a GitHub issue first, explain the lack of docs, your interest in fixing (some of) these issues, and then follow up with a PR after discussing these changes.

hi there, i launch a project, which aims to prodive API translation between Pytorch(python version) and Libtorch(C++ version).

all code lives in jupyter notebook, you can interactively write C++ in notebook.

welcome pull request.