1D Convolution where input spatial dimension isn't 2 #2358
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
sdjukicTT
requested review from
sdjordjevicTT,
svuckovicTT,
mtopalovicTT,
jserbedzijaTT and
azecevicTT
as code owners
azecevicTT
approved these changes
Mar 4, 2025
sdjordjevicTT
approved these changes
Mar 4, 2025
mrakitaTT
approved these changes
Mar 7, 2025
| @@ -17,4 +17,24 @@ module { | |||
| // CHECK: return %{{.*}} : tensor<1x1024x512xf32, #ttnn_layout3> | |||
| return %1 : tensor<1x1024x512xf32> | |||
| } | |||
|
|
|||
| func.func public @conv1d_test2(%arg0: tensor<1x7x768xbf16>, %arg1: tensor<1x192x768xbf16>) -> (tensor<1x7x768xbf16>) { | |||
There was a problem hiding this comment.
Let's just put some more descriptive test names than test1 and test2, or write some comment above each test explaining the case it is testing :)
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Outdated suggestions cannot be applied.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.
Ticket
#2357
Problem description
In TTIR to TTIR decomposition,
Legalize1DConvolutionPatternis not currently supporting spatial dimensions other than 2.What's changed
This case is already supported and the check isn't necessary. After reshaping input and output to match 2D
ttir.convolutionop and the newttir.convolutionop is picked up by theConvolutionToConv2dPattern, the dimensions are permuted and handled correctly.Added a test for this case.
Checklist