[Contributors Welcome] Add missing torch::nn layers in C++ API

Toshi · September 20, 2019, 1:16pm

I can try and work with the Normalization layer implementations.

JoshVarty · September 23, 2019, 2:23am

I would like to work on ReflectionPad1d and ReflectionPad2d

swapnilbp · October 3, 2019, 5:20pm

Hi would love to work on the Loss Functions

PyExtreme · October 8, 2019, 2:18pm

Hi ,I would like to work on CrossEntropyLoss and NLLLoss.

@yf225, please let me know in case of any conflicts.

yf225 · October 13, 2019, 12:39am

Thanks everyone for your interest! We now have many great example PRs at https://github.com/pytorch/pytorch/issues/25883, and please let us know (by responding in the Github issue) if there is any layer you would like to work on.

arso · January 10, 2020, 4:39am

I would go with MXNet over PyTorch even though when using Python I prefer PyTorch. MXNet is open to having language bindings in multiple languages. Once you get a minimum viable product you could donate it to the main project. Bindings that are part of such a large well known project are likely to attract more contributors than bindings maintained under your account for PyTorch (I could not see PyTorch officially supporting Rust).