remove CompressionParameter

[AlexanderDokuchaev]

Changes

Removed CompressionParameter

Reason for changes

Unused

Related tickets

Tests

[Copilot]

Pull request overview

This PR removes the CompressionParameter abstraction from the PyTorch backend and eliminates the now-unused compression_lr_multiplier field from quantizer specs/config serialization, simplifying quantizer parameter handling and related tests.

Changes:

• Deleted CompressionParameter from nncf.torch.layer_utils and replaced its uses with plain torch.nn.Parameter.
• Removed compression_lr_multiplier from PTQuantizerSpec constructor/state, and from min-max backend quantizer construction.
• Updated serialization test fixture to reflect the new quantizer config schema; updated mypy exclude list accordingly.

Reviewed changes

Copilot reviewed 5 out of 5 changed files in this pull request and generated no comments.

Show a summary per file

File	Description
tests/torch/function_hook/test_serialization.py	Updates expected quantizer module config (drops compression_lr_multiplier).
src/nncf/torch/quantization/layers.py	Removes CompressionParameter usage and drops compression_lr_multiplier from PTQuantizerSpec.
src/nncf/torch/layer_utils.py	Removes the CompressionParameter class (leaves only StatefulModuleInterface).
src/nncf/quantization/algorithms/min_max/torch_backend.py	Stops passing compression_lr_multiplier when creating PTQuantizerSpec.
pyproject.toml	Removes src/nncf/torch/layer_utils.py from mypy’s exclude list.