Andreyan/awq extention

[andreyanufr]

Changes

Extended AWQ algorithms for patterns Act->MatMul and Act->Multiply->MatMul with insertion for extra scales after activation.

Reason for changes

Support AWQ for wider family of LLMs

Related tickets

CVS-141131

Tests

Added unit tests

[ljaljushkin]

Could you post a job number with install and weight compression tests? Just to make sure it doesn't lead to degradation.

[codecov]

Codecov Report

Attention: Patch coverage is 98.52941% with 1 line in your changes missing coverage. Please review.

Project coverage is 91.18%. Comparing base (90a7f0d) to head (829087f).
Report is 24 commits behind head on develop.

Additional details and impacted files

@@             Coverage Diff              @@
##           develop    #2708       +/-   ##
============================================
+ Coverage    47.12%   91.18%   +44.06%     
============================================
  Files          479      483        +4     
  Lines        46072    46412      +340     
============================================
+ Hits         21712    42323    +20611     
+ Misses       24360     4089    -20271     

Files	Coverage Δ
...tion/algorithms/weight_compression/awq_patterns.py	100.00% <100.00%> (ø)
...ntization/algorithms/weight_compression/backend.py	100.00% <100.00%> (ø)
.../algorithms/weight_compression/openvino_backend.py	98.85% <100.00%> (+0.14%)	⬆️
.../quantization/algorithms/weight_compression/awq.py	91.01% <96.55%> (+0.21%)	⬆️

... and 339 files with indirect coverage changes

Flag	Coverage Δ
COMMON	41.87% <0.00%> (-1.70%)	⬇️
ONNX	34.15% <0.00%> (-0.60%)	⬇️
OPENVINO	40.91% <98.52%> (+0.97%)	⬆️
TENSORFLOW	29.39% <0.00%> (?)
TORCH	65.36% <5.88%> (?)

Flags with carried forward coverage won't be shown. Click here to find out more.

Components	Coverage Δ
common	93.55% <ø> (+23.93%)	⬆️
torch	93.65% <ø> (+60.47%)	⬆️
tensorflow	93.26% <ø> (+93.26%)	⬆️
onnx	93.06% <ø> (+<0.01%)	⬆️
openvino	94.51% <100.00%> (+0.35%)	⬆️
ptq	90.50% <98.52%> (+11.09%)	⬆️

[alexsu52]

LGTM