Torch.jl VS Clang.jl

For frontends Python has quite some head-start. In principle it would be possible to write Julia frond-ends to existing ML libraries (written e.g. in C), for example https://github.com/FluxML/Torch.jl , but the advantages over Python frontends would be very limited. Only a front-to-back Julia implementation leverages most of the language advantages like composibility and flexibility.

PyTorch for example is a C++ library with a Python user interface, see e.g. the language shares in GitHub (https://github.com/pytorch/pytorch ). There is also a Julia binding for Torch (https://github.com/FluxML/Torch.jl), but I do not know how up-to-date it is.

Are there solid C interfaces that can be used?

A large part of why I started using Julia is because calling into other languages through the C FFI is pretty easy and efficient. Most of the wrappers are a single line. If there is not existing driver support, I would pass the C headers through Clang.jl, which automatically wraps the C API in a C header.

https://github.com/JuliaInterop/Clang.jl

I most recently did this with libtiff. Here is the Clang.jl code to generate the bindings. It's less than 30 lines of sterotypical code.

https://github.com/mkitti/LibTIFF.jl/tree/main/gen

The generated bindings with a few tweaks is here:

https://github.com/mkitti/LibTIFF.jl/blob/main/src/LibTIFF.j...

If you are interested in C++ interop you can also have a look at Clang.jl and CxxWrap.jl (the usual Julia package chaos applies, where the package mentioned in old talks and docs that you find on google is superseded by some others...)