Will Tebbutt
Will Tebbutt
An interface package between MLJ and the AbstractGPs ecosystem is in the process of being registered in the general registry, so I would like to also register the package with...
```julia julia> Zygote.gradient(x -> (() -> x[:y])(), Dict(:y => 0.4)) (nothing,) ``` the gradient w.r.t. the `y` element of `x` should be `1`. This bug doesn't occur with the equivalent...
As pointed out [here](https://github.com/JuliaDiff/ChainRules.jl/pull/335) by @mcabbott , the following happens: ```julia julia> using FillArrays, Zygote julia> Zygote.gradient(x -> x[1], Fill(3, 2))[1] 2-element Array{Int64,1}: 1 0 ``` The type is innappropriate...
Zygote's current [map](https://github.com/FluxML/Zygote.jl/blob/a2f7c9731279b8ad80322c35b3a0325f03392c0e/src/lib/array.jl#L180) implementation is arguably a bit optimistic about the types of things that it's able to handle. For example, [this issue](https://github.com/JuliaGaussianProcesses/KernelFunctions.jl/issues/113) in KernelFunctions.jl cropped up because we define...
I'm trying to make the basic features of Stheno play nicely with Soss, and struggling to figure out what I'm doing wrong. In particular, here's some code that I'm trying...
Said examples are outdated and need some work to bring them back up to date, and to incorporate them into CI so that they don't degrade in the future. Not...
Currently, it's not possible to run Stheno + Turing together on 1.7 because Turing itself doesn't run on 1.7. See this PR: https://github.com/TuringLang/Turing.jl
I'm opening this to scope out + plan the best manner in which Stheno.jl can be extended (or, rather, re-extended -- we used to have functionality that did this but...
Relates to [this issue](https://github.com/HamletWantToCode/GPFlux.jl/issues/2). A MWE an be found in the `examples` directory on [this branch](https://github.com/willtebbutt/Stheno.jl/tree/wct/flux-nkn-integration). It contains a basic demo of - how to compose a `Flux` model with...
Ensuring that Stheno supports all of its functionality on the GPU (where appropriate) is quite important moving forwards. This will have to be [CuArrays.jl](https://github.com/JuliaGPU/CuArrays.jl/)-based. To achieve this, we must ensure...