You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Hello,
It looks like this still depends on theano, although it also looks like you've been revising the code to depend on aesara instead. Is there a reason it hasn't been switched to depend on aesara yet?
The text was updated successfully, but these errors were encountered:
The change of the backend to Aesara is worthy of a major version bump. At the same time we are overhauling the inner workings of PyMC3, which is why we're making this switch with "v4".
👉 PyMC3 version 3.x will remain with theano-pymc as the backend
👉 PyMC3 version 4.x will use aesara as the backend
The master branch of the PyMC3 repo already uses Aesara, but we're not making a release from it. The v4 development branch however will be merged into master soon, and then we'll also start making v4.x releases.
You can read more on the Timeline wiki page.
Thanks for the response; hoping there's a 4.x release soon. The reason I write is I'm a software engineer trying to help a scientist make a working environment for their code (with optimised libraries and other requirements) and it looks like it'll be easier to do this with aesara than the older theano codebase.
Hello,
It looks like this still depends on theano, although it also looks like you've been revising the code to depend on aesara instead. Is there a reason it hasn't been switched to depend on aesara yet?
The text was updated successfully, but these errors were encountered: