Memory increase for course?

I am hoping to use RStudio Cloud for a statistics course I will be teaching in January. The class relies heavily on rstan (and in particular brms). Unfortunately, the current 1GB memory limit in RStudio Cloud means it is impossible to compile stan models.

Is there any way to increase the memory limit for a specific project or group of users (students)? There will be fewer than 20 students in the course so it would not be a huge load.

Are these enormous models? 1Gb seems like plenty to me. Maybe there's some other problem.

The models aren't big, but Stan writes C++ programs in the background that take a lot of memory to compile.

Others have run into this problem (e.g. Error running brms model on Rstudio cloud ; brms, rstan and other packages needing compilation in RStudio Cloud ) and for now it seems like RStan is just not usable on RStudio Cloud.

1 Like

This discussion about running stan under shiny might help. I imagine it has some of the same restrictions as RStudio Cloud.

1 Like

I will look into solutions on that thread, thanks.

Though FWIW, it seems like the consensus is that it is a problem with memory:

For instance this project that runs a simple Stan example stopped working when the memory restrictions went into place: Posit Cloud

1 Like

Yes it seems to be the case. The thread I linked above shows how to compile the stan offline and then use it online to avoid this.

Ah I see. Thanks again.

This topic was automatically closed 21 days after the last reply. New replies are no longer allowed.