[PTQ] Metatypes with bias list alignment

[KodiaqQ]

Changes

• Reduced metatypes in OPERATIONS_WITH_BIAS list for ONNX;
• Reduced metatypes in OPERATIONS_WITH_BIAS list for PyTorch;
• Separated GroupConvolution from the Convolutions on ONNX;

Reason for changes

• Alignment for algorithms between backends;

Related tickets

• 133198
• 104166
• 106469

Tests

• post_training_quantization/323 - failed with regressions for ONNX (as expected)
• post_training_quantization/324 - passed (updated references)

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 84.84%. Comparing base (86ea8f6) to head (a87c1d1).

Additional details and impacted files

@@             Coverage Diff             @@
##           develop    #2506      +/-   ##
===========================================
- Coverage    91.18%   84.84%   -6.34%     
===========================================
  Files          492      492              
  Lines        45103    45122      +19     
===========================================
- Hits         41129    38286    -2843     
- Misses        3974     6836    +2862     

Files	Coverage Δ
nncf/onnx/graph/metatypes/groups.py	100.00% <ø> (ø)
nncf/onnx/graph/metatypes/onnx_metatypes.py	99.60% <100.00%> (+0.01%)	⬆️
nncf/openvino/graph/metatypes/groups.py	100.00% <ø> (ø)
nncf/torch/graph/operator_metatypes.py	99.23% <ø> (ø)

... and 56 files with indirect coverage changes

Flag	Coverage Δ
COMMON	43.93% <ø> (ø)
ONNX	34.67% <100.00%> (+0.01%)	⬆️
OPENVINO	∅ <ø> (∅)
TENSORFLOW	29.85% <0.00%> (-0.02%)	⬇️
TORCH	65.74% <0.00%> (-0.03%)	⬇️

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

Components	Coverage Δ
common	93.11% <ø> (-0.66%)	⬇️
torch	93.59% <ø> (ø)
tensorflow	93.74% <ø> (ø)
onnx	93.15% <100.00%> (+0.09%)	⬆️
openvino	25.58% <ø> (-68.49%)	⬇️
ptq	69.79% <ø> (-20.43%)	⬇️

[alexsu52]

LGTM

[KodiaqQ]

@AlexanderDokuchaev, @kshpv, please, review.