-
Various discussions have been had in various places about the correct kinds of types to implement `rrule`s for, but we've not discussed this in a central location. This problem probably occurs for som…
-
**Describe the bug**
When auto-generating an adjoint specialization, Q# reverses the order of statements, but does not reverse the order of evaluation in expressions. Because evaluation can have si…
-
I think the adjoint for the Cholesky `ldiv` is wrong :
https://github.com/FluxML/Zygote.jl/blob/56f411876d545ff758d78053d89c45b57d88e41c/src/lib/array.jl#L473
The NamedTuple uses the wrong names, …
-
I find that `hvcat` with sparse matrices is orders of magnitude slower than expected when adjoints of sparse matrices are involved. Here is an example:
```Julia
julia> using LinearAlgebra
julia> …
-
I'm using the recent master branch and the code from the docs returns:
```julia
(tmp) pkg> st
Status `/tmp/Project.toml`
[052768ef] CUDA v2.6.3
[7a1cc6ca] FFTW v1.4.1
[efe261a4] NFFT…
-
Hey, is this package intended to solve general optimal control problems?
I really wish I could use NLOptControl.jl but it [hasn't been updated to Julia 1.0](https://github.com/JuliaMPC/NLOptControl.…
-
@dsyme this merge #235 is slightly worrying. Let's discuss when we meet next time.
About the `Tensor0` case: I understand that you have some very good reason for this based on how you feel the code…
-
In https://www.tandfonline.com/doi/pdf/10.1080/10652469.2016.1164156, closed-form expressions were derived for the derivatives of Bessel functions with respect to the order. Currently these are not im…
-
We are definitely going to support optimization of real-valued functions defined on manifolds, so computing gradients of such functions is quite important. What we have now is:
- Manopt.jl seems to…
-
## Description
@SteveBronder found that softmax isn't quite working right. The reverse mode softmax code here sets the adjoint instead of incrementing it: https://github.com/stan-dev/math/blob/deve…