Thanks this worked indeed! I didn’t spot this because the ODE solver still worked and only SciML gave the problem.
I probably also had old versions of the packages, but after updating I get the same error if I don’t fix the tolerances to Float32.
1 Like