Enable dynamic shape tests for sdpa with kv cache#4067
Closed
kimishpatel wants to merge 4 commits intogh/kimishpatel/58/basefrom
Closed
Enable dynamic shape tests for sdpa with kv cache#4067kimishpatel wants to merge 4 commits intogh/kimishpatel/58/basefrom
kimishpatel wants to merge 4 commits intogh/kimishpatel/58/basefrom
Conversation
Note that in practice during prefill we will always see start_pos=0 regardless prompt size. However, the test added here also simulate batch (along seq dim) inference done not just at the beginning. This is useful for speculative decoding where you want the larger model to run batched inference for efficiency. Differential Revision: [D58874163](https://our.internmc.facebook.com/intern/diff/D58874163/) [ghstack-poisoned]
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/4067
Note: Links to docs will display an error until the docs builds have been completed. ✅ No FailuresAs of commit 4de8093 with merge base 38046ba ( This comment was automatically generated by Dr. CI and updates every 15 minutes. |
Contributor
|
This pull request was exported from Phabricator. Differential Revision: D58874163 |
This was referenced Jun 25, 2024
larryliu0820
approved these changes
Jun 26, 2024
Note that in practice during prefill we will always see start_pos=0 regardless prompt size. However, the test added here also simulate batch (along seq dim) inference done not just at the beginning. This is useful for speculative decoding where you want the larger model to run batched inference for efficiency. Differential Revision: [D58874163](https://our.internmc.facebook.com/intern/diff/D58874163/) [ghstack-poisoned]
Contributor
|
This pull request was exported from Phabricator. Differential Revision: D58874163 |
Note that in practice during prefill we will always see start_pos=0 regardless prompt size. However, the test added here also simulate batch (along seq dim) inference done not just at the beginning. This is useful for speculative decoding where you want the larger model to run batched inference for efficiency. Differential Revision: [D58874163](https://our.internmc.facebook.com/intern/diff/D58874163/) [ghstack-poisoned]
Contributor
|
This pull request was exported from Phabricator. Differential Revision: D58874163 |
Note that in practice during prefill we will always see start_pos=0 regardless prompt size. However, the test added here also simulate batch (along seq dim) inference done not just at the beginning. This is useful for speculative decoding where you want the larger model to run batched inference for efficiency. Differential Revision: [D58874163](https://our.internmc.facebook.com/intern/diff/D58874163/) [ghstack-poisoned]
Contributor
|
This pull request was exported from Phabricator. Differential Revision: D58874163 |
Contributor
|
This pull request has been merged in 38cff09. |
kedarnath03
pushed a commit
to kedarnath03/executorch
that referenced
this pull request
Jun 25, 2025
Pull Request resolved: pytorch/executorch#4067 Note that in practice during prefill we will always see start_pos=0 regardless prompt size. However, the test added here also simulate batch (along seq dim) inference done not just at the beginning. This is useful for speculative decoding where you want the larger model to run batched inference for efficiency. //unrelated failures @bypass-github-export-checks ghstack-source-id: 231807014 @exported-using-ghexport Differential Revision: [D58874163](https://our.internmc.facebook.com/intern/diff/D58874163/)
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
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.
Stack from ghstack (oldest at bottom):
Note that in practice during prefill we will always see start_pos=0 regardless
prompt size. However, the test added here also simulate batch (along seq dim)
inference done not just at the beginning. This is useful for speculative
decoding where you want the larger model to run batched inference for
efficiency.
Differential Revision: D58874163