TransformerBlock: support attention skips #14826
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
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/14826
Note: Links to docs will display an error until the docs builds have been completed. ❗ 1 Active SEVsThere are 1 currently active SEVs. If your PR is affected, please view them below: ❌ 3 New Failures, 1 Unrelated FailureAs of commit 976b081 with merge base 91f1769 ( NEW FAILURES - The following jobs have failed:
BROKEN TRUNK - The following job failed but were present on the merge base:👉 Rebase onto the `viable/strict` branch to avoid these failures
This comment was automatically generated by Dr. CI and updates every 15 minutes. |
meta-cla
bot
added
the
CLA Signed
sxu
added a commit
to sxu/executorch
that referenced
this pull request
Oct 6, 2025
Summary: We want to support attention skips. This diff modifies `TransformerBlock` to make `attention_norm` and `attention` optional. Since our export script directly constructs the `TransformerBlock`s themselves, this is enough for our use case. The top level `Transformer` class still require a single `attention_type`, to make that interface also support attention skip (which requires different configuration for each layer) is not within the scope of this diff. Differential Revision: D84003431
billmguo
approved these changes
Oct 7, 2025
jackzhxng
reviewed
Oct 7, 2025
sxu
added a commit
to sxu/executorch
that referenced
this pull request
Oct 7, 2025
Summary: We want to support attention skips. This diff modifies `TransformerBlock` to make `attention_norm` and `attention` optional. Since our export script directly constructs the `TransformerBlock`s themselves, this is enough for our use case. The top level `Transformer` class still require a single `attention_type`, to make that interface also support attention skip (which requires different configuration for each layer) is not within the scope of this diff. Reviewed By: billmguo Differential Revision: D84003431
sxu
added a commit
to sxu/executorch
that referenced
this pull request
Oct 7, 2025
Summary: We want to support attention skips. This diff modifies `TransformerBlock` to make `attention_norm` and `attention` optional. Since our export script directly constructs the `TransformerBlock`s themselves, this is enough for our use case. The top level `Transformer` class still require a single `attention_type`, to make that interface also support attention skip (which requires different configuration for each layer) is not within the scope of this diff. Reviewed By: billmguo Differential Revision: D84003431
sxu
added a commit
to sxu/executorch
that referenced
this pull request
Oct 7, 2025
Summary: We want to support attention skips. This diff modifies `TransformerBlock` to make `attention_norm` and `attention` optional. Since our export script directly constructs the `TransformerBlock`s themselves, this is enough for our use case. The top level `Transformer` class still require a single `attention_type`, to make that interface also support attention skip (which requires different configuration for each layer) is not within the scope of this diff. Reviewed By: billmguo Differential Revision: D84003431
Summary: We want to support attention skips. This diff modifies `TransformerBlock` to make `attention_norm` and `attention` optional. Since our export script directly constructs the `TransformerBlock`s themselves, this is enough for our use case. The top level `Transformer` class still require a single `attention_type`, to make that interface also support attention skip (which requires different configuration for each layer) is not within the scope of this diff. Reviewed By: billmguo Differential Revision: D84003431
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.
Summary: We want to support attention skips. This diff modifies
TransformerBlockto makeattention_normandattentionoptional. Since our export script directly constructs theTransformerBlocks themselves, this is enough for our use case. The top levelTransformerclass still require a singleattention_type, to make that interface also support attention skip (which requires different configuration for each layer) is not within the scope of this diff.Differential Revision: D84003431