Docs/vi posterior notebook#1819
Conversation
Codecov Report✅ All modified and coverable lines are covered by tests. Additional details and impacted files@@ Coverage Diff @@
## main #1819 +/- ##
=======================================
Coverage 87.86% 87.86%
=======================================
Files 140 140
Lines 12726 12726
=======================================
Hits 11182 11182
Misses 1544 1544
Flags with carried forward coverage won't be shown. Click here to find out more. |
|
Hi @Shamik-07, apologies for the delay in reviewing this! We've had a high volume of PRs recently. On a first look, it's a good idea to add a how-to-guide on VI methods. I will add a more concrete review soon. Thanks for your patience 🙏 |
|
Thanks for putting this together, and please excuse that we did not follow through with the detailed review we promised. Since this PR was opened, the VI API and the how-to-guide structure have changed substantially, and the current proposal would require a significant rewrite and revalidation rather than incremental review. We’re therefore going to close this PR while keeping #1765 open. The proposed outline remains useful, but any renewed work should start from current main, include an executed analytic validation example. |
Created a new VI notebook and updated the docs.
Note
The
qdetails have been retrieved from here and the vi_method details from hereImportant
With my limited understanding of sampling methods and with research, i have included the "when to use VI over MCMC".
Closes #1765 .