Add mean and covariance to the MvNormal variations #120
Conversation
|
Are you able to take this over and fix it to your hearts content? These things are gettingb above my git paygrade |
test/others.jl
Outdated
| @test cov(d1) ≈ cov(d2) rtol=1e-6 | ||
|
|
||
| A = randn(10) | ||
| C = A * A' + I |
There was a problem hiding this comment.
It would be nice to have testing for a larger category of matrices.
I applied the suggestions (one can just commit them in Github's webinterface, next to the comment with the suggestion). There's some minor issue with ReverseDiff and Skellam in the linked PR (which contains the fixes for the CI). I haven't merged it yet since I am waiting for @mohamed82008's suggestions on how to fix this problem. Let's see if it will be fixed within the next few days - otherwise I'll extract the CI fixes and merge them in a separate PR. |
|
Thanks so much! |
|
while we are at it, would it be ok to also add the |
|
It seems there was some problem with merging master. |
|
@devmotion I tried to rebase to see if the CI is working now, but my gitfoo is weak. This might be something you can more efficiently do. Hope that I didn't mess things up. If you completely drop this PR and just redo your change to a new one I wouldn't be offended. Sorry if I made things worse. |
|
No worries, I'll have a look and fix the git issues. I'm sorry, somehow I forgot this PR. |
|
No apologies at all, you are doing so much. Sorry I couldn't be more useful. If you do decide to throw this out and restart (and it is a trivial addition), maybe add in the variance that @nmheim suggested? |
I agree, we should add it regardless of how the git issues are resolved 🙂 |
|
Thanks a lot @jlperla! |
No description provided.