Alignment of OV and ONNX models outputs naming (#7691)

* Alignment of OV and ONNX models outputs naming

* Python tests adaptation to new naming rules

* New output naming rules

* Output name retrieval adaptation (tensor iterator node)

* Copying of tensor names during output replacement

* Multiout multinode subgraphs handling in the importer

* Proper replacement tensor naming

* Model zoo test runner adaptation

* Backwards compatible python tests runner adaptation

* If node adaptation

* Adaptation to changes in master

* Deprecation warning suppression

* Imports fix in compatibility tests

* If node adaptation to the new naming

* MaxPool python tests re-enabled

* ONNX Identity elimination adaptation

* XFAIL for the Identity op test

* Support for Param->Result models and indentity op

* Fix of the ONNX Indentity handling

* The test that fails only on windows temporarily disabled

* ONNX tensor names test adaptation

* Code cleanup

* Code formatting

* Obsolete helper removal

* One more spot where output name helper should be used

* PyApi fix for tensors with multiple names

* Don't set friendly names for unnamed ONNX nodes

* Revert "Don't set friendly names for unnamed ONNX nodes"

This reverts commit 92c7ac59b5.

* Missing dot...

* And now the mypy nonsense...

* Use get_any_name in Loop

* New way of naming result nodes in ONNX FE

src/bindings/python/src/compatibility/pyngraph/util.cpp

@@ -6,6 +6,7 @@
 
 #include <pybind11/numpy.h>
 
+#include "ngraph/op/result.hpp"
 #include "ngraph/validation_util.hpp"
 #include "ngraph/version.hpp"
 

src/bindings/python/src/openvino/runtime/ie_api.py

@@ -10,6 +10,7 @@ from openvino.pyopenvino import Core as CoreBase
 from openvino.pyopenvino import ExecutableNetwork as ExecutableNetworkBase
 from openvino.pyopenvino import InferRequest as InferRequestBase
 from openvino.pyopenvino import AsyncInferQueue as AsyncInferQueueBase
+from openvino.pyopenvino import Output
 from openvino.pyopenvino import Tensor
 from openvino.pyopenvino import Variant as VariantBase
 
@@ -42,8 +43,15 @@ def normalize_inputs(py_dict: dict, py_types: dict) -> dict:
 
 
 def get_input_types(obj: Union[InferRequestBase, ExecutableNetworkBase]) -> dict:
-    """Get all precisions from object inputs."""
-    return {i.get_any_name(): i.get_element_type() for i in obj.inputs}
+    """Map all tensor names of all inputs to the data types of those tensors."""
+
+    def map_tensor_names_to_types(input: Output) -> dict:
+        return {n: input.get_element_type() for n in input.get_names()}
+
+    input_types: dict = {}
+    for input in obj.inputs:
+        input_types = {**input_types, **map_tensor_names_to_types(input)}
+    return input_types
 
 
 class InferRequest(InferRequestBase):

src/bindings/python/tests/test_frontend/test_frontend_onnx_editor.py

@@ -142,7 +142,7 @@ def create_test_onnx_models():
         make_tensor_value_info("out4", onnx.TensorProto.FLOAT, (2, 2)),
     ]
     expected_split = onnx.helper.make_node("Split", inputs=["out1/placeholder_port_0"],
-                                           outputs=["out1", "out2"])
+                                           outputs=["out1", "out2"], name="split1", axis=0)
     expected_mul = onnx.helper.make_node("Mul", inputs=["out4/placeholder_port_0", "out4/placeholder_port_1"],
                                          outputs=["out4"])
     graph = make_graph([expected_split, expected_mul], "test_graph", input_tensors, output_tensors)
@@ -201,7 +201,7 @@ def create_test_onnx_models():
         make_tensor_value_info("out4", onnx.TensorProto.FLOAT, (2, 2)),
     ]
     expected_split = onnx.helper.make_node("Split", inputs=["out1/placeholder_port_0"],
-                                           outputs=["out1", "out2"])
+                                           outputs=["out1", "out2"], name="split1", axis=0)
     expected_mul = onnx.helper.make_node("Mul", inputs=["out4/placeholder_port_0", "out4/placeholder_port_1"],
                                          outputs=["out4"])
     graph = make_graph([expected_split, relu, expected_mul], "test_graph", input_tensors, output_tensors)

src/bindings/python/tests/test_ngraph/test_ops.py

@@ -11,7 +11,6 @@ from openvino.runtime.impl.op import Constant, Parameter
 from tests.runtime import get_runtime
 
 
-
 def binary_op(op_str, a, b):
 
     if op_str == "+":

src/bindings/python/tests_compatibility/__init__.py

@@ -144,4 +144,3 @@ xfail_issue_63136 = xfail_test(reason="Unsupported operation: CastLike")
 xfail_issue_63137 = xfail_test(reason="Unsupported operations: OptionalHasElement, OptionalGetElement")
 xfail_issue_63138 = xfail_test(reason="Missing ONNX Shape-15 support")
 xfail_issue_63643 = xfail_test(reason="RuntimeError: Unsupported operation of type: Convolution name")
-xfail_issue_54663 = xfail_test(reason="Disabled until MaxPool-8 is supported on CPU")

src/bindings/python/tests_compatibility/test_ngraph/test_ops.py

@@ -10,8 +10,6 @@ from ngraph.impl import AxisSet, Function, Shape, Type
 from ngraph.impl.op import Constant, Parameter
 from tests_compatibility.runtime import get_runtime
 
-from tests_compatibility import xfail_issue_54663
-
 
 def binary_op(op_str, a, b):