nsiccha
nsiccha
@aseyboldt that was me :+1: I performed the same experiment that you did, just for a standard normal target and for varying numbers of dimensions. From my experiment and from...
@aseyboldt > I think more important is what happens in real world posteriors. Absolutely. I've also realized that the dependence on the dimension is lower than I thought, or rather...
[Discussion on Stan Slack](https://mc-stan.slack.com/archives/C7THPDGSV/p1673291031177289) for any of the Julia folks not in the Stan Slack.
I'm just here to write that I'm having the same problems (with Julia 1.9.0).
Yes, more or less exactly. > a warmup stage that does this In the linked implementation, I'm kinda hijacking (out of laziness) the mass matrix adaptation stage to also do...
I never use bayesplot, ggplot or even R, but I can have a look some time in Feb or March. Unless Noa wants to implement it :)
Maybe something like [this](https://github.com/FedeClaudi/Term.jl/issues/264) would be nice to support?
Forgot the version: Julia version 1.10.8 + Enzyme version 0.13.28
This is mainly reshuffling some internal logic, isn't it?
I think for a simpler MWE, I actually did observe the same returned draw(s), but different post-pathfinder rng states. For the original problem, I'm unsure whether both the draws/distributions and...