Mohamed Tarek

Results 132 issues of Mohamed Tarek

Here is an example of differentiating QR decomposition using this package. Would be nice to have an example for each factorization. ```julia using LinearAlgebra, ImplicitDifferentiation, ComponentArrays, Zygote function qr_conditions(A, x)...

Hi! Here is a MWE that gives an error when trying to differentiate a function that broadcasts a FillArray of structs. ```julia using FillArrays, Zygote struct T a end f(t::T,...

needs adjoint

Turing + ReverseDiff + tape caching is currently sometimes faster than Stan but many times we are also somewhat slower especially when having loops in the model. One variable to...

good-first-issue

Zygote returns `nothing` for the gradient if the output of the function is a constant. This can happen if a uniform distribution is used alone in the model, so the...

PR #965 introduced the ability to pass user-defined variable names. The symbols of the parameters of a model is therefore no longer baked into the type of the model. This...

This line https://github.com/KristofferC/JuAFEM.jl/blob/a8d976d4e9d4bba415c3d2dbf062a75f8d814b2c/src/Dofs/DofHandler.jl#L184 will call a function that is not defined for faces with 4 corner vertices, and that have dofs on the inside of the face. This will be...

Only Float64 precision is fully supported. I will make a PR for this.

https://github.com/KristofferC/JuAFEM.jl/blob/a8d976d4e9d4bba415c3d2dbf062a75f8d814b2c/src/Dofs/ConstraintHandler.jl#L211 The function `faces` only returns tuples of corner points, so the boundary conditions on non-corner points will not be written.

`eps(ForwardDiff.Dual{Nothing}(1.0,1.0)) isa ForwardDiff.Dual` is false. Is this intentional? This is causing trouble in https://github.com/TuringLang/Turing.jl/issues/700.

We need integration testing github workflows like the ones in ChainRules and DynamicPPL to avoid breakages in TuringLang and SciML at least.