-
-
Notifications
You must be signed in to change notification settings - Fork 2k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Make VI work on v4 #4582
Make VI work on v4 #4582
Conversation
@ferrine I understand this is still WIP ? After #4579 (the rebase) replaces |
93a096d
to
c9fa127
Compare
@ferrine this was one of the last remaining PRs that still had |
I included the VI tests as a standalone job into the pytest workflow, so it's output is not contaminated by flaky tests that still exist on Running locally on Windows:
List of test failures
Hope this helps :) |
@brandonwillard I have a subset of functionality working. AEVB inference does not work and I start thinking maybe I mark this as NotImplemented?
|
I agree that we can skip that.
…On Wed, Jul 28, 2021 at 4:47 PM Maxim Kochurov ***@***.***> wrote:
@brandonwillard <https://github.com/brandonwillard> I have a subset of
functionality working. AEVB inference does not work and I start thinking
maybe I mark this as NotImplemented?
if aevb: raise NotImplementedError("AEVB is not yet supported, open an issue once you need it https://github.com/pymc-devs/pymc3/issues")
—
You are receiving this because you commented.
Reply to this email directly, view it on GitHub
<#4582 (comment)>, or
unsubscribe
<https://github.com/notifications/unsubscribe-auth/AAFETGDXT7YGZKQLTFRWAITT2AJ75ANCNFSM4Z7ATPNQ>
.
|
Hmm, the issue still appears with grouped inference that might be some shape bug/typo somewhere. I'll keep working on this. This should not be that complicated I hope |
Codecov Report
@@ Coverage Diff @@
## main #4582 +/- ##
==========================================
+ Coverage 81.44% 87.21% +5.76%
==========================================
Files 81 81
Lines 14204 14265 +61
==========================================
+ Hits 11569 12441 +872
+ Misses 2635 1824 -811
|
/pre-commit-run |
@ferrine I rebased this. |
One weird test is left and it's done. Also I remember to change skip to xfail for one test |
Tests are passing 🥳! You need to rebase from the rename, however. |
Co-authored-by: Ricardo Vieira <28983449+ricardoV94@users.noreply.github.com>
Co-authored-by: Ricardo Vieira <28983449+ricardoV94@users.noreply.github.com>
Co-authored-by: Ricardo Vieira <28983449+ricardoV94@users.noreply.github.com>
@ferrine I pushed some changes and closed outdated conversations. Can you go through the remaining open conversations and address/close them? Otherwise I think this is pretty ready to merge, and we can focus on refactoring/deprecating the remaining methods in follow up PRs |
@ricardoV94 are we good to merge? |
Only one way to find out... |
I commented one instance where detailed type info from the docstring could be annotated in the signature. |
Co-authored-by: Michael Osthege <michael.osthege@outlook.com>
/pre-commit run |
/pre-commit-run |
No it's something that should be fixed on main, I merged a PR too wrongly. |
Incredible -- congrats @ferrine and @ricardoV94! |
VI in V4
Referencing #4523 Here to work on VI module
Step 1
The first goal of this PR is to fix test errors in
because there are a lot of errors
Step 2
TBD