[CPU] OV CPU plugin fails to infer SegNext model

openvinotoolkit · Jan 23, 2025 · aeb4a2f · aeb4a2f
1 parent 8dfb9cc
commit aeb4a2f
Show file tree

Hide file tree

Showing 3 changed files with 8 additions and 8 deletions.
diff --git a/src/plugins/intel_cpu/src/transformations/cpu_opset/common/pass/convert_matmul_to_fc.cpp b/src/plugins/intel_cpu/src/transformations/cpu_opset/common/pass/convert_matmul_to_fc.cpp
@@ -69,7 +69,7 @@ ov::intel_cpu::ConvertMatMulToFC::ConvertMatMulToFC() {
          */
 
         auto get_aligned_shapes =
-            [shape_a, shape_b, rank_a, rank_b, &matmul]() -> std::tuple<bool, ov::PartialShape, ov::PartialShape> {
+            [shape_a, shape_b, rank_a, rank_b, &matmul, fc_input_a, fc_input_b]() -> std::tuple<bool, ov::PartialShape, ov::PartialShape> {
             ov::PartialShape shape_a_aligned(shape_a), shape_b_aligned(shape_b);
             size_t max_size = std::max(rank_a, rank_b);
             for (size_t i = 0, cnt = max_size - rank_a; i < cnt; ++i) {
@@ -79,10 +79,10 @@ ov::intel_cpu::ConvertMatMulToFC::ConvertMatMulToFC() {
                 shape_b_aligned.insert(shape_b_aligned.begin(), 1);
             }
 
-            if (matmul->get_transpose_a()) {
+            if (matmul->get_transpose_a() && !is_type<op::v1::Transpose>(fc_input_a.get_node())) {
                 std::swap(*(shape_a_aligned.end() - 1), *(shape_a_aligned.end() - 2));
             }
-            if (matmul->get_transpose_b()) {
+            if (matmul->get_transpose_b() && !is_type<op::v1::Transpose>(fc_input_b.get_node())) {
                 std::swap(*(shape_b_aligned.end() - 1), *(shape_b_aligned.end() - 2));
             }
 
@@ -138,12 +138,12 @@ ov::intel_cpu::ConvertMatMulToFC::ConvertMatMulToFC() {
         }
 
         // Weights normalization
-        if (!matmul->get_transpose_b()) {
+        if (!matmul->get_transpose_b() && !is_type<op::v1::Transpose>(fc_input_b.get_node())) {
             fc_input_b = create_transpose(fc_input_b, matmul->get_friendly_name() + "/transpose_b");
         }
 
         // Input normalization
-        if (matmul->get_transpose_a()) {
+        if (matmul->get_transpose_a() && !is_type<op::v1::Transpose>(fc_input_a.get_node())) {
             fc_input_a = create_transpose(fc_input_a, matmul->get_friendly_name() + "/transpose_a");
         }
 

diff --git a/src/plugins/intel_cpu/src/transformations/transformation_pipeline.cpp b/src/plugins/intel_cpu/src/transformations/transformation_pipeline.cpp
@@ -427,8 +427,8 @@ void Transformations::PreLpt(const std::vector<ov::element::Type>& defaultPrecis
         manager,
         [](const_node_ptr& node) -> bool {
             const auto consumers = node->get_output_target_inputs(0);
-            return std::all_of(consumers.begin(), consumers.end(), [](const ov::Input<ov::Node>& consumer) {
-                return !ov::is_type<ov::op::v0::MatMul>(consumer.get_node());
+            return !std::all_of(consumers.begin(), consumers.end(), [](const ov::Input<ov::Node>& consumer) {
+                return ov::is_type<ov::op::v0::MatMul>(consumer.get_node());
             });
         },
         ov::pass::KeepConstAndDecompression);

diff --git a/...intel_cpu/tests/functional/custom/subgraph_tests/src/common/matmul_decompress_convert.cpp b/...intel_cpu/tests/functional/custom/subgraph_tests/src/common/matmul_decompress_convert.cpp
@@ -503,7 +503,7 @@ namespace {
 
 const auto testParams2D_FP16_2_smoke =
     ::testing::Combine(::testing::Values(static_shapes_to_test_representation({{2, 3}, {2, 3}, {3, 4}})),
-                       ::testing::Values(std::pair<bool, bool>{false, true}),
+                       ::testing::Values(std::pair<bool, bool>{false, true}, std::pair<bool, bool>{true, false}),
                        ::testing::Values(ElementType::f16),
                        ::testing::Values(emptyConfig),
                        ::testing::ValuesIn(filter_specific_params(false)));