[NPU]Change NPUBackend log (#27073)

### Details:
[log.error of
NPUbackend](https://github.com/openvinotoolkit/openvino/blob/master/src/plugins/intel_npu/src/plugin/src/backends.cpp#L130)
will confuse the user, thought NPU backend does not impact the
compilation and just impact the inference stage.
Now [in the inference stage in compiledmodel
part](https://github.com/openvinotoolkit/openvino/pull/27073/files#diff-74bc81bb7b258118f04e81468e3ec3b05e65e714546d32246bae45eb892f6abcR125-R130),
will get a log.error output when no npu device is checked.


### Tickets:
 - 153439

Author: DanLiu2Intel

src/plugins/intel_npu/src/plugin/src/backends.cpp

@@ -111,7 +111,7 @@ NPUBackends::NPUBackends(const std::vector<AvailableBackends>& backendRegistry,
         } catch (const std::exception& ex) {
             _logger.warning("Got an error during backend '%s' loading : %s", backendName.c_str(), ex.what());
         } catch (...) {
-            _logger.error("Got an unknown error during backend '%s' loading", backendName.c_str());
+            _logger.warning("Got an unknown error during backend '%s' loading", backendName.c_str());
         }
     }
 
@@ -127,7 +127,8 @@ NPUBackends::NPUBackends(const std::vector<AvailableBackends>& backendRegistry,
     if (_backend != nullptr) {
         _logger.info("Use '%s' backend for inference", _backend->getName().c_str());
     } else {
-        _logger.error("Cannot find backend for inference. Make sure the device is available.");
+        _logger.warning("None of the backends were initialized successfully."
+                        "Only offline compilation can be done!");
     }
 }
 

src/plugins/intel_npu/src/plugin/src/compiled_model.cpp

@@ -24,6 +24,9 @@
 namespace {
 
 constexpr std::string_view NO_EXECUTOR_FOR_INFERENCE =
+    "Can't create infer request due to create executor failed! Only exports can be made.";
+
+constexpr std::string_view NO_EXECUTOR_FOR_INFERENCE_NODEVICE =
     "Can't create infer request!\n"
     "Please make sure that the device is available. Only exports can be made.";
 
@@ -118,8 +121,14 @@ std::shared_ptr<ov::IAsyncInferRequest> CompiledModel::create_infer_request() co
     if (_executorPtr == nullptr && _device != nullptr) {
         _executorPtr = _device->createExecutor(_networkPtr, _config);
     }
+
     if (_executorPtr == nullptr) {
-        OPENVINO_THROW(NO_EXECUTOR_FOR_INFERENCE);
+        if (_device != nullptr) {
+            OPENVINO_THROW(NO_EXECUTOR_FOR_INFERENCE);
+        } else {
+            _logger.error("Can not find device!");
+            OPENVINO_THROW(NO_EXECUTOR_FOR_INFERENCE_NODEVICE);
+        }
     }
 
     const std::shared_ptr<SyncInferRequest>& syncInferRequest =