Compilation of brm model never completes

Modeling
,
1

Hello!

I am trying to fit a simple Bayesian regression model in R version 4.0.2. but the message stays in “Compiling Stan program…” and never proceeds to the outcomes. I know that brm models are slower compared to other models but I am not able to get any outcomes; I was even waiting for hours. To your reference, I use Windows 10 and a computer with an 8 GB memory, while there is plenty of free space in my hard disk (more than 200 GB). The code is the following:

library(rstan)
Sys.setenv(LOCAL_CPPFLAGS = ‘-march=native -mtune=native’)
rstan_options(auto_write=TRUE)
options (mc.cores=parallel::detectCores ())
library(brms)
prior <- set_prior(“student_t(3,0,2.5)”, class = “b”)
fit <- brm(response ~ contrast + syllable + english.use + syllable : contrast + (1 | subject), data = test, family = bernoulli(“logit”), prior = prior, sample_prior = “yes”, control = list(adapt_delta = 0.95))
Compiling Stan program…

I would really appreciate it if you could give me a solution on this issue.

2

@paul.buerkner might know of a regression, but have you tried this on another computer and could you provide a minimum example that we could try out?

3

Thank you for your reply.
I found a solution for this issue. Actually, the latest version of rstan does not allow the compilation of brm models. You have to return to a previous version, that is, 2.19.3.