a short summary of questions and answers from today’s discussion:
-
folks using the LOO package on Windows - @jonah would like your help testing this PR: simplify the normalization in gpdfit by avehtari · Pull Request #187 · stan-dev/loo · GitHub
which seems to be failing for mysterious multi-processing/threading behavoirs -
the posteriordb crew (@mans_magnusson and @avehtari et al) released version 0.3, now with KTs new more complex posteriors
-
@jsocolar is playing around with Stan’s adaptation algorithm, specifically the dual averaging used for stepsize adaptation and needs help/pointers/instructions on how to build a custom version of Stan for experimentation
-
which systems which have implemented dynamic HMC in Python and R using Stan as the backend for gradients and log density?
-
@WardBrian is working on the depreciation proposal: A Deprecation Schedule for Stan