Open jolespin opened 5 years ago
Once PyMC4 is more mature we'll definitely try to port this.
Since PyMC4 devel started during TensorFlow 1.0 era does that mean the current PyMC4 needs to be rewritten?
We're targeting TF2.
When will TF2 be?
No TF2 plans scheduled
Hi all, just wanted to check in on this again, since TF2 has been officially released for some time now, is there a plan to port to TF2?
Having just bought and started the book I gotta say 1000 times this - it's a very well-written book, but it urgently needs to be updated and when it is, it should be directly to PyMC4, ideally with a brief nod / introduction to TF Probability as well. I will buy this again immediately and happily pay the full price again when we get an update. Please consider doing this soon - it's a very relevant topic and a great time right now. I would be delighted if I could get an updated version this year.
And if you should have code beforehand (e.g. a pymc4 branch) I'd be happy to alpha test it - just let us know when you start this.
Coming here to see if this has changed! Love your way of writing and would be really insightful how the TFP concepts translate to more modern syntax.
First off...thank you SO SO MUCH for updating this code and porting it to some modern tools. This a huge resource for the community and especially someone like me who had a single class in Bayesian Inference (all with manual calculations on paper) a few years ago and few resources for (re-)learning both bayesian inference concepts and a new api.
I know that the TFP port is pretty new but I was wondering if there were any plans in the future to update the code for TensorFlow 2.0 since the syntax is a little different? (I would be frustrated if someone asked me this question so my sincerest apologies)
I also heard that PyMC4 will be using TFP and TensorFlow 2.0. Will that be expected to be ported as well? (Again, I know that is a HUGE task but just thought I'd knock out 2 birds with one stone in this 2 part question).