Sebastian Micluța-Câmpeanu
Sebastian Micluța-Câmpeanu
Hi, thanks for creating this package. I was wondering if it would be possible for LocalRegistry to create PRs when registering packages so that it can be used in conjunction...
Attempting to use symbolic variable names from MTK models (e.g. `ODESystem`) in `save_idxs` errors with invalid index of type `Num`. To get around this I used ```julia indexof(sym, syms) =...
From slack: George Datseris:juliapool: Nov 6th at 7:00 PM So we have a method that integrates many many initial conditions and maps the trajectories into a low-dimensional vector of ~10...
MWE ```julia using Symbolics Symbolics.derivative(exp(im*x),x) ``` gives ``` ERROR: MethodError: no method matching decompose(::Num) Closest candidates are: decompose(::Integer) at float.jl:555 decompose(::Float16) at float.jl:557 decompose(::Float32) at float.jl:568 ... Stacktrace: [1] isfinite(x::Num)...
Would it make sense to be able to autodiff through `knn`s? I tried using `ForwardDiff` ```julia using NearestNeighbors using ForwardDiff data = rand(3, 10^3) kdtree = KDTree(data) ForwardDiff.gradient(x->sum(knn(kdtree, x, 4)[2]),...
It would be useful to be able to use `OptimizationBBO` without having all the output printed to stdout. In order to do this, we can use the `TraceMode` keyword argument,...
It would be nice to be able to use `MonteCarloSummary` and the rest of the analysis tools for functions of the solutions.
I was thinking that it may be useful to consider an option to toggle progress monitoring, as during benchmarks it may not always be useful. An other thing that may...
I think it would be useful to discuss here the design changes required for working on GPUs. @jgreener64, could you detail a bit on what's needed for working on GPUs?...
I've looked at the performance of a standard liquid argon simulation and I'd like to propose some small design changes. - [x] The interactions could be stored in tuples or...