-
### 🐛 Describe the bug
In my application, I need to take the nth order mixed derivative of a function. However, I found that the torch.autograd.grad computation time increases exponentially as n incr…
-
[Automatic Differentiation and SciML: What Can Go Wrong](https://www.youtube.com/watch?v=OyFP565kDUI) | Chris Rackauckas
00:00:00 Welcome
00:00:45 Content outline
00:01:50 Prologue: Why do differ…
-
After much digging and testing, it appears you are currently not supporting sparse matrix constructions for AD. Is this correct? In particular, I am in need of a MvNormal distribution that is specifi…
-
This is annoying if you run into it. I imagine that the best fix is a special method for the `size` of this complicated type of `Complex` array.
_
_ _ _(_)_ | …
-
I have a few questions on how to improve the implementations for the upcoming varmat and OpenCL PRs and make them more maintainable. This would also cleanup how we do the transform MIR with OpenCL.
…
-
Is this already captured in UQ section?
-
Today I was looking at https://github.com/conal/agda-cat-linear with @conal , and I was working on getting the Nix-based build working. It uses the current master versions of the `standard-library` an…
-
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,…
-
A few comments for consideration in the matrix autodiff design document.
1. As much as the matrix of vars enables better performance it also greatly facilities the implementation matrix-valued reve…
-
```julia
julia> using LinearAlgebra, BenchmarkTools
julia> A = rand(200, 10);
julia> B = rand(200, 200);
julia> @btime $A' * $B * $A;
45.817 μs (2 allocations: 16.62 KiB)
julia> @btime…