Skip to content

Automatically tie kv cache i/o quantizers in AimetQuantization#2261

Merged
jambayk merged 1 commit intomicrosoft:mainfrom
CodeLinaro:dev/mtuttle/add_kv_io_tying_to_aimet
Nov 17, 2025
Merged

Automatically tie kv cache i/o quantizers in AimetQuantization#2261
jambayk merged 1 commit intomicrosoft:mainfrom
CodeLinaro:dev/mtuttle/add_kv_io_tying_to_aimet

Conversation

@michaelgtuttle
Copy link
Contributor

Describe your changes

Automatically aligns quantization parameters (scale, offset, bitwidth) between kv cache input and output tensors in AimetQuantization pass.

Checklist before requesting a review

  • Add unit tests for this change.
  • Make sure all tests can pass.
  • Update documents if necessary.
  • Lint and apply fixes to your code by running lintrunner -a
  • Is this a user-facing change? If yes, give a description of this change to be included in the release notes.

(Optional) Issue link

Signed-off-by: Michael Tuttle <mtuttle@qti.qualcomm.com>
@jambayk jambayk enabled auto-merge (squash) November 17, 2025 22:31
@jambayk jambayk merged commit 29a5aee into microsoft:main Nov 17, 2025
15 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants