[AWQ] Expand scale dims to match activation dims #3746
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
github-actions
bot
added
NNCF OpenVINO
daniil-lyakhov
force-pushed
the
dl/awq_batch_dim
branch
3 times, most recently
from
6e829b5 to
3bb01f7
Compare
daniil-lyakhov
commented
Nov 20, 2025
| merge_weight = self._backend_entity.get_weight(merge_node, port_id, model, graph) | ||
| merge_weight = (merge_weight * a_scale).astype(weight_dtype) | ||
| self._backend_entity.set_weight(merge_node, port_id, model, graph, merge_weight) | ||
| a_scale = fns.transpose(a_scale) |
Collaborator
Author
There was a problem hiding this comment.
As this a_scale is always being squeezed before the usage, there is no need to do transpose here
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
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Outdated suggestions cannot be applied.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.
Changes
AWQ: multiply scale shape is expanded to match the activation shape length: Mamba models have an AWQ pattern which has 3 dims, and insertion of default 2d scale leads to an error during inference
Reason for changes
To support AWQ algo for mamba models
Related tickets
173277
Tests
tests/cross_fw/test_templates/template_test_weights_compression.py::test_awq_scale_reference is updated to test the non mergable AWQ branch, testing the branch + new reshape implementation