[DO NOT MERGE] Simplifying the convtranspose decomposition #411
Conversation
This file contains hidden or 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
chaitanyakamarapu
changed the title
jorickert
approved these changes
Aug 13, 2025
jorickert
left a comment
jorickert
left a comment
There was a problem hiding this comment.
I did not verify the math behind it, but it looks technically fine.
Just make sure that we have a check that batch == 1
ehsan-toosi
approved these changes
Aug 13, 2025
ehsan-toosi
left a comment
ehsan-toosi
left a comment
There was a problem hiding this comment.
The decomposition looks good to me but I have one question. Has it ever checked the convTranspose is 2d? I didn't see that check or maybe I overlooked it.
Comment on lines
+30
to
+33
| // CHECK: %[[VAL_21:.*]] = "onnx.Slice"(%[[VAL_20]], %[[VAL_11]], %[[VAL_10]], %[[VAL_13]], %[[VAL_12]]) : (tensor<256x512x6x6xf32>, tensor<2xi64>, tensor<2xi64>, tensor<2xi64>, tensor<2xi64>) -> tensor<256x512x3x3xf32> | ||
| // CHECK: %[[VAL_22:.*]] = "onnx.Slice"(%[[VAL_20]], %[[VAL_9]], %[[VAL_8]], %[[VAL_13]], %[[VAL_12]]) : (tensor<256x512x6x6xf32>, tensor<2xi64>, tensor<2xi64>, tensor<2xi64>, tensor<2xi64>) -> tensor<256x512x3x3xf32> | ||
| // CHECK: %[[VAL_23:.*]] = "onnx.Slice"(%[[VAL_20]], %[[VAL_7]], %[[VAL_6]], %[[VAL_13]], %[[VAL_12]]) : (tensor<256x512x6x6xf32>, tensor<2xi64>, tensor<2xi64>, tensor<2xi64>, tensor<2xi64>) -> tensor<256x512x3x3xf32> | ||
| // CHECK: %[[VAL_24:.*]] = "onnx.Slice"(%[[VAL_20]], %[[VAL_5]], %[[VAL_4]], %[[VAL_13]], %[[VAL_12]]) : (tensor<256x512x6x6xf32>, tensor<2xi64>, tensor<2xi64>, tensor<2xi64>, tensor<2xi64>) -> tensor<256x512x3x3xf32> |
There was a problem hiding this comment.
Could conv transpose have output some channel not multiple of 8?
…into chaitany.simply_convtransose_decomposition
chaitanyakamarapu
changed the title
Collaborator
Author
We have a check for this. |
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.
Simplifying the four phase conv merge in convtranspose phased decomposition