loo

loo copied to clipboard

testing running the vignettes on GHA

1

Testing running the vignettes on GitHub Actions. Will definitely fail on Windows but maybe not on Mac and/or Ubuntu.

jgabry

At some point I had added \eqn{}, but forgot to change R code to LaTeX equations, This fixes that

avehtari

Related to discussion in [a Stan discourse post](https://discourse.mc-stan.org/t/is-it-possible-to-get-positive-and-negative-values-from-the-loo-compare-function/34690) ROS book and some Stan discourse answers show how to make Jacobian adjustment for elpd_loo values if different models use different transformations...

avehtari

Use more functions from the posterior package

1

Release 2.7.0 imports posterior. We could use more posterior package functions. This would reduce possibility of having different implementations and thus reduce maintenance burden. Potential functions to replace - `ess_rfun`...

avehtari

Port C++ psis code to loo?

1

In Stan we have a lot of the code for psis sampling available as C++. https://github.com/stan-dev/stan/blob/develop/src/stan/services/pathfinder/psis.hpp It should be a good bit faster. How much time do you ballpark the...

SteveBronder

We have the Mix-IS-LOO vignette https://mc-stan.org/loo/articles/loo2-mixis.html, but it would be useful to add support for creating a loo object from mixture posterior `log_lik`. Assuming the Stan model code implements the...

avehtari

"Error in if (varx == 0) { : missing value where TRUE/FALSE needed" with fit$loo() but not loo::loo(fit$draws("log_lik"))

16

Hi, Unfortunately I don't have a reproducible example because this is just popping up with a few versions of a big model. When I fit the model with cmdstanr, the...

mhollanders

`neg_binomial_2_rng` can obstruct `loo_moment_match()`

2

I'm working on a Stan program that includes a `generated quantities` block that generates `log_lik` for use with `loo` and `yrep` for posterior predictive checks. This looks something like (with...

kthayashi

Removed unused fft function and relying on posterior

6

One test was failing due to minor numerical differences in algorithms: average diff: 0.00975, largest diff: 0.01472. I updated the expected test result at reference-results/relative_eff.rds. posterior::autovariance is already used. Starts...

VisruthSK

#289 bumps testthat to the third edition for snapshots. There are potentially other parts of the test suite which could be updated to use new functionality introduced in 3e. Moving...

VisruthSK