Conversation
|
Some tests are performed with a manual finite diff with specific tangents that remain in the domain |
devmotion
left a comment
There was a problem hiding this comment.
Thanks! I'll do a proper review later, here just two quick initial comments.
Codecov Report
@@ Coverage Diff @@
## master #1534 +/- ##
==========================================
+ Coverage 85.74% 85.84% +0.10%
==========================================
Files 129 129
Lines 7975 8018 +43
==========================================
+ Hits 6838 6883 +45
+ Misses 1137 1135 -2
Continue to review full report at Codecov.
|
Co-authored-by: David Widmann <devmotion@users.noreply.github.com>
Co-authored-by: David Widmann <devmotion@users.noreply.github.com>
Co-authored-by: David Widmann <devmotion@users.noreply.github.com>
Co-authored-by: David Widmann <devmotion@users.noreply.github.com>
Co-authored-by: David Widmann <devmotion@users.noreply.github.com>
|
On why |
|
For the tests, the rrule one fails even with forward finite diff, the finite diff version ends ups being NaN |
I assume the problem with finite differencing in general is that perturbing |
Co-authored-by: David Widmann <devmotion@users.noreply.github.com>
Co-authored-by: David Widmann <devmotion@users.noreply.github.com>
devmotion
left a comment
There was a problem hiding this comment.
Looks good to me, is there anything else you would like to include in this PR?
|
setting it as draft to not merge it yet, I have to sort out some things before |
|
Merging as already discussed and approved |
|
No all good on my side thanks a lot!
…On Thu, May 26, 2022, 16:03 David Widmann ***@***.***> wrote:
***@***.**** approved this pull request.
Looks good to me, is there anything else you would like to include in this
PR?
—
Reply to this email directly, view it on GitHub
<#1534 (review)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AB2FDMT3K3DHUZCJ2KONZ7DVL7KIPANCNFSM5UF5ANPA>
.
You are receiving this because you authored the thread.Message ID:
***@***.***>
|
This adds the ChainRules rules for Dirichlet, constructor and logpdf