[GPU] Add squeeze/unsqueeze to the compressed conv1x1 transformation

[mdvoretc-intel]

Details:

• Adds an activation squeeze/unsqueeze node pair to the ConvertWeightCompressedConv1x1ToMatmul transformation when outermost dimension is a static 1
• This is intended to reduce the activation to 3D, allowing compile-time weight reordering on non-systolic devices

Tickets:

• CVS-182319

AI Assistance:

• AI assistance used: no

[Copilot]

Pull request overview

Adds an activation squeeze/unsqueeze pair around the MatMul emitted by ConvertWeightCompressedConv1x1ToMatmul when the activation’s leading dimension is statically 1, aiming to reduce activation rank to 3D to enable compile-time weight reordering on non-systolic devices.

Changes:

• Add optional activation squeeze before MatMul and matching unsqueeze after MatMul in the conversion pass.
• Extend the existing transformation unit test to cover batched vs non-batched inputs and update the reference graph accordingly.

Reviewed changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated 2 comments.

File	Description
src/common/transformations/src/transformations/op_conversions/convert_weight_compressed_conv1x1_to_matmul.cpp	Implements conditional squeeze/unsqueeze reshapes around the produced MatMul.
src/common/transformations/tests/op_conversions/convert_weight_compressed_conv1x1_to_matmul_test.cpp	Expands parameterized coverage (adds with_batched_input) and updates the reference model to match the new reshape behavior.

[Copilot]

[BLOCKER] The new squeeze/unsqueeze reshape constants rely on shape_out[-1].get_length(). If the last dimension is dynamic (e.g., input has dynamic channel/feature dim), get_length() throws and the transformation will crash during pass execution. Consider building the reshape pattern without requiring static last dim (e.g., use special_zero=true with a 0 in the last position to copy the input’s last dimension), and apply the same fix to the unsqueeze block as well.

Suggested change

	auto shape_out = activation->get_output_partial_shape(0);
	auto squeeze_const =
	std::make_shared<ov::op::v0::Constant>(ov::element::i64,
	ov::Shape{3},
	std::vector<int64_t>{1, -1, shape_out[-1].get_length()});
	MatcherPass::register_new_node(squeeze_const);
	auto squeeze = std::make_shared<ov::op::v1::Reshape>(activation, squeeze_const, false);
	// Reshape [1, D1, ..., Dn] -> [1, -1, Dn] while safely copying the last dimension
	// without requiring it to be statically known.
	auto squeeze_const =
	std::make_shared<ov::op::v0::Constant>(ov::element::i64,
	ov::Shape{3},
	std::vector<int64_t>{1, -1, 0});
	MatcherPass::register_new_node(squeeze_const);
	auto squeeze = std::make_shared<ov::op::v1::Reshape>(activation, squeeze_const, true);

[Copilot]

[HIGH] The new squeeze/unsqueeze behavior is only validated with fully static shapes. Since the implementation currently depends on shape introspection (and should support dynamic channel dims once the get_length() issue is addressed), please add a regression test case where the activation’s last dimension is dynamic (or weights/output dim is dynamic) to ensure the transformation does not throw and preserves the expected output rank/shape.

[CuriousPanCake]

You're inserting Reshape, but calling it unsqueeze. Is this how it should be?

[mdvoretc-intel]

The squeeze/unsqueeze operations are represented by reshapes here to avoid adding new node types at this stage in the pipeline.