Automatic Differentiation #82
Merged
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
Jutho
reviewed
Sep 27, 2023
| return _repartition(p, N₁) | ||
| end | ||
|
|
||
| TensorKit.block(t::ZeroTangent, c::Sector) = t |
There was a problem hiding this comment.
This seems a bit suspicious? Why is this needed?
There was a problem hiding this comment.
This is mostly to avoid having to manually deal with the ZeroTangent type. For example, a QR decomposition where the final result does not depend on R, would generate a ZeroTangent for dR, which is just an abstract representation that behaves as the zero vector in any (co)vectorspace. As some of the rrules are implemented "blockwise", this would either require manually checking if a tangent is a ZeroTangent, or, which is what I chose to do, rely on the hope that the compiler would recognize that the blockwise operation results in ZeroTangents anyways, and thus automatically takes care of this
Jutho
reviewed
Sep 27, 2023
lkdvos
commented
Sep 27, 2023
tensorscalar now has a `rrule`
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Outdated suggestions cannot be applied.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.
This PR adds rrules and tests for most linear algebra methods.
Note that the tests for
tsvdare disabled when working withComplex, as I have not been able to figure out how to make the lack of uniqueness ofUandVwork with finite differences and or the ChainRulesTestUtils interface.