Allow customization of sdpa in executorch.py #38827
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
guangy10
approved these changes
Jun 16, 2025
Earlier PR put executorch specific sdpa and mask function in the export function. This prevent any customization that can be done to sdpa, prior to export. By moving this to __init__, we still keep the original behavior but allow users like optimum-executorch to override sdpa by setting model.config._attn_implementation.
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Outdated suggestions cannot be applied.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.
What does this PR do?
Earlier PR put executorch specific sdpa and mask function in the export function. This prevent any customization that can be done to sdpa, prior to export. By moving this to init, we still keep the original behavior but allow users like optimum-executorch to override sdpa by setting model.config._attn_implementation.
Fixes # (issue)
optimum-executorch to export with custom sdpa. Mainly this
optimum-cli export executorch --model "HuggingFaceTB/SmolLM2-135M" --task "text-generation" --recipe "xnnpack" --output_dir="smollm" --use_custom_sdpain optimum-executorch should work (although it needs this pr + huggingface/optimum-executorch#82Before submitting
Pull Request section?
to it if that's the case.
documentation guidelines, and
here are tips on formatting docstrings.
Who can review?
Anyone in the community is free to review the PR once the tests have passed. Feel free to tag
members/contributors who may be interested in your PR.