[Torch] Fix advanced indexing with boolean mask #13306
Merged
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
Collaborator
|
Thanks for contributing to TVM! Please refer to the contributing guidelines https://tvm.apache.org/docs/contribute/ for useful information and tips. Please request code reviews from Reviewers by @-ing them in a comment. Generated by tvm-bot |
This was referenced Nov 7, 2022
Contributor
|
Thanks @masahi |
xinetzone
pushed a commit
to daobook/tvm
that referenced
this pull request
Nov 10, 2022
* [Torch] Fix advanced indexing with boolean mask * add comment
xinetzone
pushed a commit
to daobook/tvm
that referenced
this pull request
Nov 25, 2022
* [Torch] Fix advanced indexing with boolean mask * add comment
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 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Outdated suggestions cannot be applied.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.
Fixes #12292
Fixes #13296
PyTorch uses the same op,
aten::index, to represent indexing by both integer indices and a boolean mask. In our PT converter, Relayadv_indexop is used to convertaten::index, but it doesn't correctly recognize indexing by mask - it will treat True / False as indices 0 / 1. So our converter generates an invalid model if the PT model uses such indexing, see #12292 and #13296.I fixed this issue by explicitly tuning the mask into integer indices using
argwhere, which will also take care of the inherent dynamism in indexing by mask.cc @comaniac @junrushao @shingjan @yelite