ExpandingMan

> To make this easier to review, can this be split up into three parts? I will try my best, 2 and 3 might be a little bit tricky, but...

Yeah in my opinion there's still a lot that could be done, but certainly at this point it would make far more sense to do it in a different PR,...

I have never been able to get this not to segfault. ``` System: Host: theia1 Kernel: 6.10.6-10-MANJARO arch: x86_64 bits: 64 Desktop: Qtile v: 0.28.2.dev0+gf1ed49bc.d20240813 Distro: Manjaro Linux CPU: Info:...

I thought that maybe this was something very obscure with `Revise` or something, so I ran it with ``` julia --startup-file=no scrap2.jl ``` and it still segfaults the same way.

For sake of posterity, note the following: I have so far reproduced this on two different (but very similar x86 linux 6.10 ryzen) machines. I have also tried nuking my...

Didn't validate it too deeply, but it looks ok, certainly doesn't segfault at least 👍

Alright, this looks like it works. Again, while this may be a fix this solution is not very nice and involves a lot of redundancy, but I think anything better...

This is only a problem on GPU because of scalar indexing. On current main it should give ``` ERROR: MethodError: getindex(::OneHotMatrix{UInt32, CuArray{UInt32, 1, CUDA.DeviceMemory}}, ::Int64, ::Int64) is ambiguous. Candidates: getindex(x::OneHotArray{var"#s3",...

I think there is a bigger issue here of what exactly do you do in the edge cases that this package allows where you would seem to have to scalar...

Yeah, the front-end for Enzyme has had a lot of issues (especially with StaticArrays). You should mostly find that the low level `autodiff` from Enzyme is blazing fast in most...