bgoodri
bgoodri
We are go for StanHeaders with the additional `size -> math::size` thing. I am a bit perplexed as to how StanHeaders 2.26.27 ultimately worked with survstan and all the other...
Uploading the new StanHeaders (that only changed one `size` to `math::size`) somehow broke the following rstantools-using packages on Fedora and M1: [Rglt](https://cran.r-project.org/web/checks/check_results_Rlgt.html) [bmlm](https://cran.r-project.org/web/checks/check_results_bmlm.html) [eggCounts](https://cran.r-project.org/web/checks/check_results_eggCounts.html) It is the same error that...
It is possible that we can go ahead with the rstan 2.26 because doing so would not make those three packages any "worse" as far as the CRAN checks are...
OK. Let me see if I can merge your rstan-cran branch into develop without too many merge conflicts. On Thu, Sep 7, 2023 at 12:43 PM Andrew Johnson ***@***.***> wrote:...
Hopefully just a few more hours Frank. On Thu, Sep 7, 2023, 5:44 PM Frank Harrell ***@***.***> wrote: > I appreciate the work on this. I just updated my rmsb...
@paul-buerkner @harrelfe There is now a RStan on CRAN that accepts both array syntaxes. We will have to do a follow-up RStan based on 2.26 before we can do one...
That is presumably the parser parsing. I had the same thing with rstanarm (and then it failed). So, if it is just a compile time thing (in the case of...
I think writing it somewhere when the sampler is done would be fine. On Mon, Jul 24, 2017 at 9:24 PM, Daniel Lee wrote: > @bgoodri , do you envision...
If the caller is getting the draws of everything, then it doesn't matter very much if the means and variances (or whatever else) are being accumulated because it is a...
No, after thinning, which is to say only for the post-warmup iterations that are written out. For parameters whose draws are retained, the accumulated means and variances should match what...