Closed wwiecek closed 1 year ago
Could we compute the stan models, store them via posteriordb and access them on github during the vignette compilation process? https://github.com/MansMeg/posteriordb/tree/master/rpackage That way we wouldn't need to run anything locally.
Disclaimer: I've never used this package, so I don't know if this would work.
I don't see much of an advantage in that. Or rather, I think it's maybe a solution if we were sure we wanted to store the results somewhere, but I thought we agreed that the first solution was to check if it is possible to submit precompiled Markdown packages in such a way that CRAN will not run all of the (lengthy) code. That is the only issue, I think.
AFAIK CRAN does use the precompiled doc anyway, but runs the code to check for any errors. Or something like that.
That's fair, just a passing thought when I was reading about the package.
Sure -- didn't mean to shoot it down entirely!
Currently I run some
loocv
andlogit
models withiter=500
but soon even this won't be enough. I could run and store outputs withiter = 5000
Could either store within package or at a stable online location (GitHub?)