[PT FE] Use logging instead of print (openvinotoolkit#26739)

### Details:
 - *Use logging instead of print*
 - *Improve logs*

### Tickets:
 - *CVS-136015*

src/bindings/python/src/openvino/frontend/pytorch/gptq.py

@@ -4,8 +4,12 @@
 # flake8: noqa
 # mypy: ignore-errors
 
-import torch
 from functools import partial
+import logging
+import torch
+
+log = logging.getLogger(__name__)
+
 
 # Wraps a single tensor to a module to prevent it from jit.freezing
 # It depends on a tensor dtype whether it will be preserved from freezing. Refer to the decoder code to learn which types will be preserved.
@@ -168,13 +172,18 @@ def unpatch_model(model):
                 del m._openvino_u4_compression_submodule_qweights
                 del m._openvino_u4_compression_submodule_qzeros
             except Exception as error:
-                print('[ WARNING ] Exception raised during GPTQ model unpatching. Depending on the exact issue it may lead to broken original model')
-                print(error)
+                log.warning("Exception raised during GPTQ model unpatching. "
+                            "Depending on the exact issue it may lead to broken "
+                            "original model.\n%s", error)
 
 
 def detect_gptq_model_raw(model):
-    return model and getattr(model, 'config', None) and getattr(model.config, 'quantization_config', None) and model.config.quantization_config.quant_method == 'gptq'
+    return (model and getattr(model, 'config', None) and
+            getattr(model.config, 'quantization_config', None) and
+            model.config.quantization_config.quant_method == 'gptq')
 
 
 def detect_gptq_model(model):
-    return detect_gptq_model_raw(model) or getattr(model, 'model', None) and detect_gptq_model_raw(model.model)
+    return (detect_gptq_model_raw(model) or
+            getattr(model, 'model', None) and
+            detect_gptq_model_raw(model.model))

src/bindings/python/src/openvino/frontend/pytorch/patch_model.py

@@ -4,9 +4,12 @@
 # flake8: noqa
 # mypy: ignore-errors
 
+import logging
 import torch
 from openvino.frontend.pytorch import ModuleExtension
 
+log = logging.getLogger(__name__)
+
 
 class no_jit_trace:
     def __enter__(self):
@@ -65,8 +68,10 @@ def new_forward(*args, **kwargs):
     for name, m in model.named_modules():
         if hasattr(m, orig_forward_name):
             # already patched, skipping with a warning because it is unexpected
-            print(f'[ WARNING ] Unexpectedly found already patched module {name} while applying ModuleExtension during PyTorch model conversion. '
-                  'Result of the conversion maybe broken. Depending on the exact issue it may lead to broken original model.')
+            log.warning("Unexpectedly found already patched module %s while applying "
+                        "ModuleExtension during PyTorch model conversion. "
+                        "Result of the conversion maybe broken. Depending on the exact issue "
+                        "it may lead to broken original model.", name)
             continue
         module_patcher(m, name)
 
@@ -78,9 +83,9 @@ def unpatch_model(model, orig_forward_name):
                 m.forward = getattr(m, orig_forward_name)
                 delattr(m, orig_forward_name)
             except Exception as error:
-                print('[ WARNING ] Exception raised during model unpatching. Depending on the exact issue it may lead to broken original model.')
-                print('Original exception details:')
-                print(error)
+                log.warning("Exception raised during model unpatching. "
+                            "Depending on the exact issue it may lead to broken original model.\n"
+                            "Original exception details:\n%s", error)
 
 
 def __make_16bit_traceable(model: torch.nn.Module):

src/bindings/python/src/openvino/frontend/pytorch/ts_decoder.py

@@ -16,13 +16,15 @@
     graph_has_ops,
 )
 from openvino.runtime import opset11 as ops
-from openvino.frontend.pytorch import gptq
-from openvino.frontend.pytorch import patch_model
+from openvino.frontend.pytorch import gptq, patch_model
 from openvino.frontend.pytorch.module_extension import ModuleExtension
 
+import inspect
+import logging
 import typing
 import torch
-import inspect
+
+log = logging.getLogger(__name__)
 
 
 class TorchScriptPythonDecoder(Decoder):
@@ -57,19 +59,21 @@ def __init__(
             except Exception as e:
                 if example_input is not None:
                     msg = "tracing"
-                    help_msg = "Please check correctness of provided 'example_input'. "
-                    "Sometimes models can be converted in scripted mode, please try running "
-                    "conversion without 'example_input'."
+                    help_msg = "Please check correctness of provided 'example_input'. " \
+                        "Sometimes models can be converted in scripted mode, please try running " \
+                        "conversion without 'example_input'.\n"
                 else:
                     msg = "scripting"
-                    help_msg = "\nTracing sometimes provide better results, please provide valid 'example_input' argument."
+                    help_msg = "Tracing sometimes provide better results, " \
+                        "please provide valid 'example_input' argument.\n"
                 raise RuntimeError(
-                    f"Couldn't get TorchScript module by {msg}. With exception:\n{e}\n{help_msg} "
+                    f"Couldn't get TorchScript module by {msg}.\n{help_msg} "
                     "You can also provide TorchScript module that you obtained"
                     " yourself, please refer to PyTorch documentation: "
                     "https://pytorch.org/tutorials/beginner/Intro_to_TorchScript_tutorial.html."
-                )
+                ) from e
             self.graph_element = pt_module.inlined_graph
+            log.debug("Inlined graph:\n%s", pt_module.inlined_graph)
             self.alias_db = self.graph_element.alias_db()
         else:
             self.graph_element = graph_element
@@ -121,7 +125,8 @@ def _get_scripted_model(self, pt_module, example_inputs=None, skip_freeze=False)
             if example_inputs is None:
                 if self.module_extensions:
                     raise RuntimeError(
-                        "ModuleExtension is not supported for scripting. Please provide valid example_input argument to run tracing.")
+                        "ModuleExtension is not supported for scripting. "
+                        "Please provide valid example_input argument to run tracing.")
                 scripted = torch.jit.script(pt_module)
                 freeze_by_default = True
             else:
@@ -140,10 +145,10 @@ def _get_scripted_model(self, pt_module, example_inputs=None, skip_freeze=False)
                         gptq.patch_model(pt_module)
                         gptq_patched = True
                     except Exception as error:
-                        print(
-                            "[ WARNING ] Failed patching of AutoGPTQ model. Error message:\n", error)
-                        print(
-                            "[ WARNING ] Tracing of the model will likely be unsuccessful or incorrect")
+                        log.warning(
+                            "Failed patching of AutoGPTQ model. Error message:\n%s"
+                            "\nTracing of the model will likely be unsuccessful or incorrect",
+                            error)
                         gptq.unpatch_model(pt_module)
                         gptq_patched = False
 

src/frontends/pytorch/src/translate_session.cpp

@@ -211,7 +211,7 @@ std::shared_ptr<Model> TranslateSession::convert_pytorch_model(
                                         fw_tensor_id);
 
 #ifdef ENABLE_OPENVINO_DEBUG
-                const auto out_type = context.get_output_type(i);
+                const auto out_type = simplified_type_interpret(context.get_output_type(i));
                 if (out_type.is<element::Type>()) {
                     if (!converted_outputs[i].get_element_type().compatible(out_type.as<element::Type>())) {
                         OPENVINO_DEBUG("[WARNING] Produced output type for operation ",