Floating point error using adjoint_sensitivities
when tspan
includes 16f0
#1133
Labels
adjoint_sensitivities
when tspan
includes 16f0
#1133
Describe the bug 🐞
Adjoint solve immediately fails with a floating point precision error exactly at 16f0, regardless of the ODEFunction, solver, etc.
Expected behavior
The solve should not fail, or at least its propensity to fail is smooth in
tspan
Minimal Reproducible Example 👇
Error & Stacktrace⚠️
In the above MRE, a warning is thrown and the gradients are identically zero due to the adjoint solve failing.
Environment (please complete the following information):
using Pkg; Pkg.status()
using Pkg; Pkg.status(; mode = PKGMODE_MANIFEST)
versioninfo()
Additional context
Add any other context about the problem here.
The text was updated successfully, but these errors were encountered: