[PTQ][MinMax][Torch] One shared quantizer is used for all unified scale quantization points #2622
Conversation
This file contains 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 PT
daniil-lyakhov
force-pushed
the
dl/torch/min_max_unified_scales_support
branch
from
fe0f86f to
c36b461
Compare
Codecov ReportAttention: Patch coverage is
Additional details and impacted files@@ Coverage Diff @@
## develop #2622 +/- ##
============================================
- Coverage 91.16% 77.93% -13.23%
============================================
Files 494 494
Lines 45373 45401 +28
============================================
- Hits 41363 35383 -5980
- Misses 4010 10018 +6008
... and 102 files with indirect coverage changes
Flags with carried forward coverage won't be shown. Click here to find out more.
|
KodiaqQ
reviewed
Apr 12, 2024
KodiaqQ
approved these changes
Apr 12, 2024
AlexanderDokuchaev
requested changes
Apr 12, 2024
Co-authored-by: Alexander Dokuchaev <[email protected]>
AlexanderDokuchaev
approved these changes
Apr 12, 2024
KodiaqQ
added a commit
that referenced
this pull request
Aug 9, 2024
### Changes - Fixed the `MinMax` algorithm for the `FakeConvert` insertion case. ### Reason for changes - Incorrect `MinMax` behaviour introduced in #2622. ### Related tickets - 149128 ### Tests - Manual Incorrect behaviour - two identical layers were inserted:  Corrected behaviour - only one layer for each branch was inserted: 
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Outdated suggestions cannot be applied.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.
Changes
create_unified_scales_quantizers_insertion_commandsis introduced: it receives several target points and one quantization parameter. Depending on implementation, one or several insertion commands are generated and returned back to the common algorithm.Reason for changes
PTSharedFNInsertionCommandto make quantizers aligned during QAT in comparison with OV/ONNX backend, which can use separate commands/quantizers for each insertion point without any restrictionsRelated tickets
104304
Tests
[Template test] test_ptq_params: test_unified_scales_command_creation
test_create_shared_quantizer_insertion_command
Jobs
manual/job/post_training_quantization/350/: passed