Document the interaction behaviour of @grad
and compile
.
#250
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
TLDR:
This PR adds documentation to
@grad
that specifies what the@grad
ed function should return, and documents that any variables defined outside of the returned adjoint function have their valuesfrozen
to their compiled ones.Context:
In this discussion from issue #243 I fumbled my way through working out how
@grad
functions in relation tocompile
. Unsurprisingly to anyone who understands adjoints, I found out that any intermediate variables defined in the body of the function could be used but not updated when the gradient is compiled withcompile
.This PR is just a bit of documentation added to the unexported macro
@grad
so that anyone else who dives into ReverseDiff might have an easier time concretely understanding the expected effects ofcompile
.Any suggestions on changes to language or wording are more than welcome.