[Relay][Bugfix] fix the wrong calculate logic of operator flip in PyTorch frontend#15752
Merged
Conversation
github-actions
Bot
requested review from
Hzfengsy,
MasterJH5574 and
echuraev
jikechao
changed the title
echuraev
reviewed
Sep 15, 2023
echuraev
left a comment
echuraev
left a comment
Contributor
There was a problem hiding this comment.
In general LGTM. Thanks for fixing bugs in TVM! Small comment.
Co-authored-by: Egor Churaev <egor.churaev@gmail.com>
MasterJH5574
approved these changes
Sep 16, 2023
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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
4 participants
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
The original implementation of Flip in PyTorch converter mistaken the type of attribute
axisin the Flip operator as an integer. Thus, It only parses the first element of theaxisand will give a wrong calculation result when the length ofaxisis more than one. According to the PyTorch documentation here, the type ofaxisis a list or tuple.This PR corrected the incorrect implementation of the algorithm of
torch.flipconverter and added a regression test.cc @echuraev @Hzfengsy @MasterJH5574