-
-
Notifications
You must be signed in to change notification settings - Fork 155
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
Use Numba's Generator
support
#1245
Use Numba's Generator
support
#1245
Conversation
This comment was marked as resolved.
This comment was marked as resolved.
8b03ded
to
da9f069
Compare
c111b34
to
f7b5dca
Compare
Codecov Report
Additional details and impacted files@@ Coverage Diff @@
## main #1245 +/- ##
=======================================
Coverage 74.12% 74.12%
=======================================
Files 174 174
Lines 48654 48654
Branches 10353 10353
=======================================
Hits 36064 36064
Misses 10300 10300
Partials 2290 2290 |
68c1716
to
40fa538
Compare
OK, this should be good merge once the Numba 0.57 RC release is created containing all the |
40fa538
to
3bb1dca
Compare
3bb1dca
to
77c077a
Compare
77c077a
to
31b77a2
Compare
8a83740
to
9304396
Compare
c608e10
to
61bd864
Compare
450f759
to
611077f
Compare
The current failure will be fixed by #1496. |
611077f
to
4d1c477
Compare
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
This looks good to me, thanks! Just to note, there are a few codecov warnings.
Closes #814.
dirichlet
isn't currently supported in Numba.gumbel
isn't currently supported in Numba.binomial
isn't currently supported in Numba.Object-mode is being used. This will need a follow-up issue for a non-object-mode implementation.
hypergeometric
isn't currently supported in Numba.Object-mode is being used. This will need a follow-up issue for a non-object-mode implementation.
scan
runs performed without merging produce slightly different sample results between Numba and Python modes.This might be due to the extra RNG state
copy
s in the PythonRandomVariable.perform
implementations, but I haven't checked yet.Update: It was due to the
copy
s, and they've been added to the Numba implementations; however, they require the use of object-mode, which is somewhat undesirable. This will need a follow-up issue for a non-object-mode implementation.