non mutating gradient #159
Draft
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| @@ -130,7 +130,12 @@ function finite_difference_gradient( | |||
| end | |||
| end | |||
| cache = GradientCache(df, x, fdtype, returntype, inplace) | |||
| finite_difference_gradient!(df, f, x, cache, relstep=relstep, absstep=absstep, dir=dir) | |||
| if typeof(x) <: AbstractArray && fdtype == Val(:central) | |||
There was a problem hiding this comment.
why the check for AbstractArray? I would think that the non-! call should always do a non-mutating version by default.
There was a problem hiding this comment.
The check for AbstractArray is only because I changed the vector->scalar code and not the scalar->vector code.
Since by #157 , the latter should be replaced with jacobian calls.
Once that is done, x should be restrained to Num.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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Outdated suggestions cannot be applied.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.
Prototype of a version that is
Zygote(nested differentiation) andStaticArrayscompatible to fix #148 .This works as long as the input element type and the output type are the same.
Could use some advice on how to convert the element type of an arbitrary array to the correct output type.
Quite a bit of performance might be sacrificed for this generalization.
This approach differs from how this is handled in
ForwardDiff, where there is a special dispatch forStaticArrays, and it is assumed that any otherAbstractArrayis able to handle mutation or gets transformed into a similar mutable array, andZygoteis (or should be) handled by a custom rule.