SciML / OrdinaryDiffEq.jl

High performance ordinary differential equation (ODE) and differential-algebraic equation (DAE) solvers, including neural ordinary differential equations (neural ODEs) and scientific machine learning (SciML)
https://diffeq.sciml.ai/latest/
Other
521 stars 198 forks source link

feat: allow initialization of null integrators #2228

Open AayushSabharwal opened 1 month ago

AayushSabharwal commented 1 month ago

Requires

Checklist

Additional context

Add any other context about the problem here.

ChrisRackauckas commented 1 week ago

What's a case that needs this?

AayushSabharwal commented 1 week ago

This was part of the rework that allows parameters to be unknowns in the initialization system. That bit is on hold for the time being, pending the discussion in https://github.com/SciML/ModelingToolkit.jl/issues/2665

ChrisRackauckas commented 1 week ago

Don't you just need to add the initializationprob call up here? https://github.com/SciML/DiffEqBase.jl/blob/master/src/solve.jl#L678

AayushSabharwal commented 1 week ago

Hmm yeah that does look better. I'll try that, thanks!