search

SciML / QuasiMonteCarlo.jl

Lightweight and easy generation of quasi-Monte Carlo sequences with a ton of different methods on one API for easy parameter exploration in scientific machine learning (SciML)
https://docs.sciml.ai/QuasiMonteCarlo/stable/
MIT License
103 stars 26 forks source link

Lazy iteration over points #68

Open ParadaCarleton opened 1 year ago

ParadaCarleton commented 1 year ago

It's generally a lot better for performance to iterate lazily over QMC points, especially when the number of points is quite large. When users need the whole set, there's nothing stopping them from calling collect.

ChrisRackauckas commented 1 year ago

Agreed. It would be nice to have an alternative to sample which returns an iterator for this.

ChrisRackauckas commented 1 year ago

Is https://github.com/SciML/QuasiMonteCarlo.jl/pull/85 sufficient @ParadaCarleton ?

dmetivie commented 1 year ago

I think what @ParadaCarleton had in mind is not for randomizing (which #85 do) but rather when sampling. Like this is actually done for Sobol points in Sobol.jl with the next! function.