Ancestors

Written by Nick Tierney on 2024-11-13 at 00:30

Super stoked to announce that {greta} version 0.5.0 is now on CRAN! Thank you to Nick Golding for making this release happen, and for his continued support, I learnt a lot!

Some features I'm really excited about:

• Now uses Tensorflow 2.0 (which involved a heavy refactor of internals)

• Installation of python dependencies has been revamped - it is closer than ever to "just working".

• Some new print methods for greta arrays and MCMC, these will only get better.

You can read the full news below:

https://github.com/greta-dev/greta/releases/tag/v0.5.0

[#]rstats

=> More informations about this toot | More toots from njtierney@aus.social

Toot

Written by Matthew Kay on 2024-11-13 at 01:00

Cool!! Been meaning to check this out for awhile and investigate integration with {posterior} and {ggdist}

=> More informations about this toot | More toots from mjskay.com@bsky.brid.gy

Descendants

Written by Nick Tierney on 2024-11-13 at 01:03

@mjskay.com Amazing! I'd be really interested in your thoughts on {greta}, I really think it is a super nice way to define models. There's a lot of fertile ground for future development in this space, I think. I haven't played around much with posterior, lately I've been more on the package dev end of things. The last time I had to wrangle MCMC output I wrote https://github.com/njtierney/mmcc to help manage it, it looks like posterior does a lot of what I wanted, I think?

=> More informations about this toot | More toots from njtierney@aus.social

Written by Matthew Kay on 2024-11-13 at 01:14

For sure! been meaning to reach out to you and other folks working on packages with objects representing distributions... my main role in {posterior} is pushing posterior::rvar, and I've had some involvement in {distributional} (mostly from kicking the tires while rewriting {ggdist} to use it)...

=> More informations about this toot | More toots from mjskay.com@bsky.brid.gy

Written by Matthew Kay on 2024-11-13 at 01:16

... but my secret / not-so-secret goal is some unified interface for distribution representation in R such that people can get all the way from model fitting to visualization without leaving a probabilistic programming abstraction. I think this roughly works with posterior::rvar() + {ggdist} but...

=> More informations about this toot | More toots from mjskay.com@bsky.brid.gy