Mohamed Tarek

I will look into the speed issue of `map` and `mapreduce` in master.

Thanks for your response. I will be presenting my review of existing open source topology optimization software (nearly 20) and my own WIP framework https://github.com/mohamed82008/TopOpt.jl in WCSMO-13. The main ones...

I am able to yes. I will do it in the weekend unless someone wants to beat me to it. It's basically exactly this implementation https://discourse.julialang.org/t/zygote-jl-how-to-get-the-gradient-of-sparse-matrix/59067/12 plus unit tests.

https://github.com/JuliaDiff/ChainRules.jl/pull/796

Note that GalacticOptim is currently undergoing some refactoring to break it down into sub-packages in the same repo.

And that's still too heavy 🤔, interesting.

Also try running the inference in the https://github.com/mohamed82008/DynamicPPL_NeurIPS environment. I think I fixed some bugs there for dynamic programs.

Yes activate the environment and run your code there.

Any update here?

You may find this discussion useful https://discourse.julialang.org/t/optimizing-a-physics-experiment/93434/3.