-
Notifications
You must be signed in to change notification settings - Fork 20
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
Fix nan gradients in TraceEnum_ELBO when evaluated eagerly #544
Conversation
Why is this blocked by #545? |
#545 is needed to unblock |
Why do we need to unblock the |
Sorry for making it confusing, I wasn't sure how to organize these PRs. I will make a separate PR for |
No worries, I just don't want your substantive work to be held up with yet more adjoint/alpha-conversion insanity. |
Do the Pyro |
Yes, just tested it.
|
Fixes nan gradients in
TraceEnum_ELBO
when evaluated eagerly (mentioned in #493).The problem seems to be that
reflect.interpret
in line 62 leads to substitutions that haveadjoint
as abase_interpretation
which in turn leads to some extra expressions being added to thetape
. Here I fix it by changing interpretation to_old_interpretation
.This also solvestest_adjoint.py::test_sequential_sum_product_adjoint
:xfail_param(MarkovProduct, reason="mysteriously doubles adjoint values?")
(blocked by #545 ).