Generalize canonicalization pattern for more aten.sub/div/mul/add op #1209
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
Vremold
changed the title
ZihengJiang
changed the title
silvasean
requested changes
Aug 11, 2022
There was a problem hiding this comment.
I see a lot of repeated code in this PR. Can we have a function like this which handles all ops? I think it would make it easier to share all the code and would avoid the templates
LogicalResult rewrite0DTensorOp(Operation *op, ...) {
// dyn_cast `op` to determine the operation.
}
ff2a957 to
4bffccd
Compare
Done! |
silvasean
approved these changes
Aug 15, 2022
Vremold
force-pushed
the
generalize-cononicalization-pattern
branch
from
4bffccd to
4e24e08
Compare
|
Merged. Thanks! @silvasean |
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Outdated suggestions cannot be applied.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 generalizes the canonicalization pattern in PR #935 to fit more AtenOps (like
AtenSubTensorOp,AtenMulScalarOp, andAtenDivTensorModeOp). The basic idea is eliminating these tensor ops when their tensor operands are zero-ranked tensors, which come fromValueTensorLiteralOporPrimNumToTensorScalarOp, and replacing them with corresponding scalar ops.This PR also helps to solve issue #1182 .