PyMC3 vs PyMC4 for a new project?
maresb opened this issue · 2 comments
Greetings! I hope this is an appropriate place to ask this... I'm fairly new to Bayesian modelling, and I would like to use it for some real-world projects. On one hand I'm very impressed with the extensive PyMC3 documentation, but on the other hand having TFP as a backend seems so much more powerful.
I'm leaning towards starting with PyMC3 because of the pre-release disclaimer:
Do not use for anything serious.
I have been watching PyMC4 for a while and have always seen a lot of activity, yet I'm unable to find any indication for whether the rough timescale for the first release version would be closer to a month or a year. If PyMC4 is on the verge of release, I'd much rather invest my time into learning that. I would be grateful for any advice. Thanks so much for your time!
Please post general questions to the Discourse site. The issues are for reporting bugs only. Thanks.
Sorry for the noise, and thanks so much for the Discourse reference! Just in case someone else stumbles across this, my question is more-or-less addressed here:
https://discourse.pymc.io/t/is-pymc4-dead-or-just-resting/5330