Add measurements postprocessing script#2260
Merged
Merged
Conversation
Signed-off-by: Artur Kloniecki <arturx.kloniecki@intel.com>
Signed-off-by: Artur Kloniecki <arturx.kloniecki@intel.com>
|
The code quality check failed, please run |
|
The docs for this PR live here. All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update. |
Collaborator
|
@AKloniecki @astachowiczhabana So this PR replaces Adam's previous one (#2238), right? |
Signed-off-by: Artur Kloniecki <arturx.kloniecki@intel.com>
Collaborator
Author
yes in deed. |
pbielak
approved these changes
Sep 16, 2025
regisss
approved these changes
Sep 16, 2025
Collaborator
regisss
left a comment
regisss
left a comment
There was a problem hiding this comment.
LGTM!
Let me know when I should merge
Collaborator
Author
|
We are ready for merge, please cherry-pick the change to v1.20-release as mentioned in the PR description as well. |
regisss
pushed a commit
that referenced
this pull request
Sep 16, 2025
Signed-off-by: Artur Kloniecki <arturx.kloniecki@intel.com>
Collaborator
Done |
astachowiczhabana
pushed a commit
that referenced
this pull request
Sep 17, 2025
Signed-off-by: Artur Kloniecki <arturx.kloniecki@intel.com>
gplutop7
pushed a commit
to HabanaAI/optimum-habana-fork
that referenced
this pull request
Oct 15, 2025
…e#685) Signed-off-by: Artur Kloniecki <arturx.kloniecki@intel.com> Co-authored-by: Artur KlonieckiX <arturx.kloniecki@intel.com>
gplutop7
pushed a commit
to HabanaAI/optimum-habana-fork
that referenced
this pull request
Nov 6, 2025
…e#685) Signed-off-by: Artur Kloniecki <arturx.kloniecki@intel.com> Co-authored-by: Artur KlonieckiX <arturx.kloniecki@intel.com>
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.
What does this PR do?
This PR adds a convenience proxy script that calls a measurements' postprocessing method from Intel Neural-Compressor library.
This script aligns the scales for matmul_av, matmul_qk with those for k_cache and v_cache, to avoid dequant-quant-scale op before matmul_qk and matmul_av.
This leads to better performance on larger batch sizes and context lengths.
README.md has been updated to add execution of the script after measurements are done.
This change should be cherry-picked to v1.20-release branch.