JingYu Ning

`Conv` is not working for Complex when using CUDA

5

```julia using Flux using CUDA CUDA.allowscalar(false) # T = Float32 T = ComplexF32 m = Chain( Conv((3, ), 1=>2, pad=1), ) |> gpu # 10 points 1 channel with batchsize=2...

Enhance transform

1

* [chebyshev](https://github.com/CliMA/OperatorFlux.jl/blob/main/src/transform_chebyshev.jl) * [legendre](https://github.com/CliMA/OperatorFlux.jl/blob/main/src/transform_legendre.jl) * [multiwavelet](https://arxiv.org/abs/2109.13459) [Ref](https://github.com/CliMA/OperatorFlux.jl/tree/main/src)

Extend `exp(A::AbstractMatrix)`

3

The method `exp(A::AbstractMatrix)` is too limited and doesn't support most of the lazy types: ```julia julia> using InfiniteArrays julia> a = cache(Ones(∞, ∞)); a[1, 1] = 0; b = view(a,...

ref: [implementation in `gdown`](https://github.com/wkentaro/gdown/pull/90)

This package needs to be fully tested and patched on Windows. We found some known issues listed [here](https://github.com/foldfelis/Terming.jl#state-of-supporting-windows).