Andraž Jelinčič

> I think should be totally fine -- it can go into the `dense_info` output of a solver step. Take a look at how Runge--Kutta methods output the intermediate stages,...

@patrick-kidger I now made a temporary fix, but there are two things that remain to be solved: 1. For some reason `eqx.filter_eval_shape(term.vf, 0.0, y0, args)` doesn't work when `term` is...

Great news @patrick-kidger @lockwo: what we discussed worked! Thanks for your advice! Patrick, if there is nothing else about the code itself that you'd like me to change, then I'll...

Thanks so much for the review, Patrick! And sorry for all my dummy mistakes 😅. I made all the smaller edits already and tomorrow I'll write a short example, a...

I added a brief Langevin example, a backwards in time test and fixed up the docs. I also did another check of the code, so hopefully there's no more rogue...

That's a good point, I wanted to keep it short due to what Patrick said, but if he agrees I'm happy to make it longer. Also it seems you have...

Thanks so much Owen! Not sure if this is useful, but in case you'd like to see the Bayesian logistic regression code, it's here: https://github.com/andyElking/diffrax_STLA/blob/devel/mcmc/bayes_log_reg.ipynb

Two things to mention: - Even the Euler discretisation of ULD has negative Lyapunov exponents (i.e. two paths with different initial conditions but same BM will converge together over time)...

Let's continue over email.

Hi Patrick, I made all the fixes you recommended. I went through the code again today and I think I don't see any more issues anywhere (but you'll probably still...