This is an archive of the discontinued LLVM Phabricator instance.

Bugfix for grouped convolution op

Authored by gpetters94 on Jun 29 2022, 11:15 PM.



Corrects the output order for Conv2dNgchwFgchw to the order expected by Torch-MLIR.

Diff Detail

Event Timeline

gpetters94 created this revision.Jun 29 2022, 11:15 PM
gpetters94 requested review of this revision.Jun 29 2022, 11:15 PM
silvas added a comment.Jul 6 2022, 1:25 PM

What is the bug, and how does this fix it?

gpetters94 added a comment.EditedJul 6 2022, 3:54 PM

What is the bug, and how does this fix it?

I was wrong about the order that Torch-MLIR expects for the output of this op because the test model I was using masked it. This corrects the output order to work with any grouped case, not just that specific model.

silvas accepted this revision.Jul 6 2022, 4:12 PM

Okay, can you update the commit message to reflect that? Right now it's a bit hard to understand.

This revision is now accepted and ready to land.Jul 6 2022, 4:12 PM
gpetters94 edited the summary of this revision. (Show Details)Jul 6 2022, 4:46 PM

Can do. Is this better?

This revision was automatically updated to reflect the committed changes.