Change composite names for JAX and refactor int8 #2431
Merged
anko-intel merged 8 commits intomasterfrom Apr 16, 2026
Merged
Conversation
Signed-off-by: Bartosz Kowalski <bartosz.kowalski@intel.com>
Signed-off-by: Bartosz Kowalski <bartosz.kowalski@intel.com>
bkowalskiINTEL
force-pushed
the
dev/bkowalsk/jax_int8_pr
branch
from
5380709 to
a9bfaf0
Compare
bkowalskiINTEL
force-pushed
the
dev/bkowalsk/jax_int8_pr
branch
from
3bd98e2 to
3f5aaf6
Compare
bkowalskiINTEL
commented
Apr 9, 2026
Contributor
There was a problem hiding this comment.
Pull request overview
This PR standardizes JAX composite names for quantize/dequantize ops and applies an int8-focused refactor/workaround intended to improve XLA compatibility.
Changes:
- Renamed
jax.lax.compositenames for quantize/dequantize to unifiedinc.quantize/inc.dequantize. - Refactored asymmetric dequantization arithmetic and extended
get_q_params()withemulate_asymmetric; zero-point is nowint32. - Updated static-quant layers to store activation zero-points as
int32and adjusted int8/asymmetric wiring in several prepare/convert paths.
Reviewed changes
Copilot reviewed 2 out of 2 changed files in this pull request and generated 4 comments.
| File | Description |
|---|---|
neural_compressor/jax/utils/utility.py |
Composite renames; asymmetric dequant arithmetic tweak; get_q_params() API/typing changes (incl. int32 zero-point + emulation flag). |
neural_compressor/jax/quantization/layers_static.py |
Switch activation zero-point weights to int32 and refactor int8/asymmetric paths in static quant layers. |
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
bkowalskiINTEL
commented
Apr 9, 2026
bkowalskiINTEL
commented
Apr 9, 2026
Force scales to be fp32 for int8 - temporary [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci Revert "Force scales to be fp32 for int8 - temporary" This reverts commit 3f5aaf6. Code cleanup Signed-off-by: Bartosz Kowalski <bartosz.kowalski@intel.com>
bkowalskiINTEL
force-pushed
the
dev/bkowalsk/jax_int8_pr
branch
from
33c1f69 to
052ffa5
Compare
anko-intel
force-pushed
the
dev/bkowalsk/jax_int8_pr
branch
from
47fb6e2 to
a8ed876
Compare
Signed-off-by: Andrzej Kotłowski <andrzej.kotlowski@intel.com>
anko-intel
changed the title
|
Azure Pipelines: Successfully started running 1 pipeline(s). 8 pipeline(s) were filtered out due to trigger conditions. |
1 similar comment
|
Azure Pipelines: Successfully started running 1 pipeline(s). 8 pipeline(s) were filtered out due to trigger conditions. |
|
Azure Pipelines: Successfully started running 1 pipeline(s). 8 pipeline(s) were filtered out due to trigger conditions. |
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
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.
Type of Change
others
Description
Removed suffixes from quantize/dequantize composites - inc.quantize_int8... to inc.quantize, etc.