Ziyi Yin

thunk still runs for non Flux.params which leads to unnecessary computation

2

Hello! I have a minimum example here about Flux and ChainRulesCore ```julia using Flux, ChainRulesCore import ChainRulesCore.rrule function f(a::Float32, b::Float32) return a * b end function rrule(::typeof(f), a::Float32, b::Float32) println("rrule...

Time-varying & matrix-free observation operators?

1

Hi, I am writing to kindly request an extra feature to make `H` an array of matrix-free operators, where `H[i]` is the observation operator of the `i`-th time step and...

Clean up docs for sim src, add LS-RTM examples

6

1. Clean up the sim src documentation in the `doc` folder 2. Add LS-RTM example with anti-chattering 3. Add LS-RTM example with sim src

overload DepthScaling and Topmute operators for PhysicalParameter?

1

Would it make sense to overload depth scaling and top mute operators? Now the default operation of depth scaling on physical parameter is given in `https://github.com/slimgroup/JUDI.jl/blob/f22170d3709bf9f4320190b6915d063570c7e003/src/TimeModeling/ModelStructure.jl#L216` where it directly outputs...

use `reverse` to replace deprecated `flipdim`

2

`flipdim` already deprecated so changed it to `reverse`

domain/range difference of JOLI leads to error on lsqr

3

```Julia julia> using IterativeSolvers, JOLI julia> F = joDFT(5; DDT=Float32, RDT=ComplexF32) (joLinearFunction{Float32, ComplexF32}, (5, 5), "joDFT_p") julia> b = randn(Float32, 5); julia> lsqr(F', b) ┌ Info: JOLI method error for...

The type conversion `T` seems unnecessary and will throw error especially when strict input/output types are specified for the linear operator `A`. For example, when `A` takes complex input and...

`Threads.@threads` introduces a try/catch not supported error in Zygote AD

4

Hi, Is there any reason why `Threads.@threads` would introduce a try/catch not supported error when doing AD with Zygote? Below is a minimal failing example where `f1` involves `Threads.@threads` but...

`ForwardDiff.jacobian` throws error for `fft`

1

reproducing script is below ``` julia (@v1.8) pkg> activate --temp Activating new project at `/var/folders/29/j2y0t06d1918t1tymbv45wgr0000gn/T/jl_T8WfIS` (jl_T8WfIS) pkg> add ForwardDiff, FFTW Updating registry at `~/.julia/registries/General.toml` Resolving package versions... Installed ChainRulesCore ─...