diff --git a/nncf/common/initialization/batchnorm_adaptation.py b/nncf/common/initialization/batchnorm_adaptation.py
index 8b8e2ce8d8b..f3febc0ce7c 100644
--- a/nncf/common/initialization/batchnorm_adaptation.py
+++ b/nncf/common/initialization/batchnorm_adaptation.py
@@ -12,7 +12,7 @@
 import math
 from abc import ABC
 from abc import abstractmethod
-from typing import Optional
+from typing import Optional, Type
 
 from nncf.api.compression import TModel
 from nncf.common.initialization.dataloader import NNCFDataLoader
@@ -83,6 +83,7 @@ def run(self, model: TModel) -> None:
         :param model: A model for which the algorithm will be applied.
         """
         backend = get_backend(model)
+        impl_cls: Type[BatchnormAdaptationAlgorithmImpl]
         if backend is BackendType.TORCH:
             from nncf.torch.batchnorm_adaptation import PTBatchnormAdaptationAlgorithmImpl
 
diff --git a/nncf/common/initialization/dataloader.py b/nncf/common/initialization/dataloader.py
index 07fa4b98ab9..5934121a338 100644
--- a/nncf/common/initialization/dataloader.py
+++ b/nncf/common/initialization/dataloader.py
@@ -11,6 +11,7 @@
 """Interface for user-defined data usage during the compression algorithm initialization process."""
 from abc import ABC
 from abc import abstractmethod
+from typing import Any
 
 from nncf.common.utils.api_marker import api
 
@@ -31,7 +32,7 @@ def batch_size(self) -> int:
         """
 
     @abstractmethod
-    def __iter__(self):
+    def __iter__(self) -> Any:
         """
         Creates an iterator for the elements of a custom data source.
         The returned iterator implements the Python Iterator protocol.
diff --git a/nncf/torch/extensions/__init__.py b/nncf/torch/extensions/__init__.py
index 59edea1ca24..2dcedab4da0 100644
--- a/nncf/torch/extensions/__init__.py
+++ b/nncf/torch/extensions/__init__.py
@@ -24,7 +24,6 @@
 
 import nncf
 from nncf.common.logging import nncf_logger
-from nncf.common.logging.logger import extension_is_loading_info_log
 from nncf.common.utils.api_marker import api
 from nncf.common.utils.registry import Registry
 
@@ -95,27 +94,27 @@ def get(self, fn_name: str) -> Callable:
         if self._loaded_namespace is None:
             timeout = int(os.environ.get(EXTENSION_LOAD_TIMEOUT_ENV_VAR, DEFAULT_EXTENSION_LOAD_TIMEOUT))
             timeout = timeout if timeout > 0 else None
-
-            with extension_is_loading_info_log(self._loader.name()):
-                try:
-                    pool = ThreadPool(processes=1)
-                    async_result = pool.apply_async(self._loader.load)
-                    self._loaded_namespace = async_result.get(timeout=timeout)
-                except MPTimeoutError as error:
-                    msg = textwrap.dedent(
-                        f"""\
-                        The extension load function failed to execute within {timeout} seconds.
-                        This may be due to leftover lock files from the PyTorch C++ extension build process.
-                        If this is the case, running the following command should help:
-                            rm -rf {self._loader.get_build_dir()}
-                        For a machine with poor performance, you may try increasing the time limit by setting the environment variable:
-                            {EXTENSION_LOAD_TIMEOUT_ENV_VAR}=180
-                        Or disable timeout by set:
-                            {EXTENSION_LOAD_TIMEOUT_ENV_VAR}=0
-                        For more information, see FAQ entry at: https://github.com/openvinotoolkit/nncf/blob/develop/docs/FAQ.md#importing-anything-from-nncftorch-hangs
-                        """  # noqa: E501
-                    )
-                    raise ExtensionLoaderTimeoutException(msg) from error
+            nncf_logger.info(f"Compiling and loading torch extension: {self._loader.name()}...")
+            try:
+                pool = ThreadPool(processes=1)
+                async_result = pool.apply_async(self._loader.load)
+                self._loaded_namespace = async_result.get(timeout=timeout)
+            except MPTimeoutError as error:
+                msg = textwrap.dedent(
+                    f"""\
+                    The extension load function failed to execute within {timeout} seconds.
+                    This may be due to leftover lock files from the PyTorch C++ extension build process.
+                    If this is the case, running the following command should help:
+                        rm -rf {self._loader.get_build_dir()}
+                    For a machine with poor performance, you may try increasing the time limit by setting the environment variable:
+                        {EXTENSION_LOAD_TIMEOUT_ENV_VAR}=180
+                    Or disable timeout by set:
+                        {EXTENSION_LOAD_TIMEOUT_ENV_VAR}=0
+                    For more information, see FAQ entry at: https://github.com/openvinotoolkit/nncf/blob/develop/docs/FAQ.md#importing-anything-from-nncftorch-hangs
+                    """  # noqa: E501
+                )
+                raise ExtensionLoaderTimeoutException(msg) from error
+            nncf_logger.info(f"Finished loading torch extension: {self._loader.name()}")
 
         return getattr(self._loaded_namespace, fn_name)
 
diff --git a/pyproject.toml b/pyproject.toml
index f9601087a1c..2bbfe807db4 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -107,8 +107,10 @@ files = [
     "nncf/common/stateful_classes_registry.py",
     "nncf/common/strip.py",
     "nncf/common/tensor.py",
+    "nncf/common/collector.py",
     "nncf/common/accuracy_aware_training",
     "nncf/common/graph",
+    "nncf/common/initialization",
     "nncf/common/sparsity",
     "nncf/common/tensor_statistics",
     "nncf/common/utils",