Update Python API stub files

Author: sys-openvino-ci

src/bindings/python/src/openvino/tools/ovc/convert.pyi

@@ -8,7 +8,7 @@ from openvino.tools.ovc.logger import restore_logger_state
 import openvino._ov_api
 import pathlib as pathlib
 __all__: list[str] = ['Model', 'convert_model', 'get_all_cli_parser', 'get_logger_state', 'pathlib', 'restore_logger_state']
-def convert_model(input_model: [str, pathlib.Path, typing.Any, list], input: [list, dict, str] = None, output: [str, list] = None, example_input: typing.Any = None, extension: [str, pathlib.Path, list, typing.Any] = None, verbose: bool = False, share_weights: bool = True) -> openvino._ov_api.Model:
+def convert_model(input_model: [str, pathlib.Path, typing.Any, list], input: [list, dict, str] = None, output: [str, list] = None, example_input: typing.Any = None, extension: [str, pathlib.Path, list, typing.Any] = None, verbose: bool = False, share_weights: bool = True, dynamo: bool = False) -> openvino._ov_api.Model:
     """
     
         Converts the model from original framework to OpenVino Model.
@@ -83,6 +83,15 @@ def convert_model(input_model: [str, pathlib.Path, typing.Any, list], input: [li
                 then mmap is used to allocate weights directly from file. If input model is
                 runtime object, then original memory regions allocated in the original model
                 are reused for weights in the converted model.
+            :param dynamo:
+                Export a PyTorch torch.nn.Module using torch.export instead of
+                torch.jit.trace. Requires example_input and PyTorch >= 2.6.
+                Default is False.
+                The resulting model uses static shapes derived from example_input by default.
+                To enable dynamic dimensions, combine with the input parameter:
+                dimensions set to -1 or Dimension(-1) become fully dynamic
+                (torch.export.Dim.AUTO), and bounded dimensions such as Dimension(1, 10)
+                are exported with explicit min/max constraints.
     
         Returns:
             openvino.Model

src/bindings/python/src/openvino/tools/ovc/moc_frontend/pytorch_frontend_utils.pyi

@@ -3,13 +3,35 @@ from __future__ import annotations
 from openvino._pyopenvino import PartialShape
 from openvino._pyopenvino import Tensor
 from openvino.tools.ovc.cli_parser import InputCutInfo as _InputCutInfo
+from openvino.tools.ovc.cli_parser import input_to_input_cut_info
 from openvino.tools.ovc.cli_parser import single_input_to_input_cut_info
 from openvino.tools.ovc.error import Error
 import logging as log
 import numpy as np
 import pathlib as pathlib
 import sys as sys
-__all__: list[str] = ['Error', 'PartialShape', 'Tensor', 'extract_input_info_from_example', 'extract_module_extensions', 'flatten_inputs', 'get_pytorch_decoder', 'get_pytorch_decoder_for_model_on_disk', 'get_value_from_list_or_dict', 'log', 'np', 'pathlib', 'prepare_torch_inputs', 'single_input_to_input_cut_info', 'sys', 'to_torch_tensor', 'update_list_or_dict']
+__all__: list[str] = ['Error', 'PartialShape', 'Tensor', 'extract_input_info_from_example', 'extract_module_extensions', 'flatten_inputs', 'get_pytorch_decoder', 'get_pytorch_decoder_for_model_on_disk', 'get_value_from_list_or_dict', 'input_to_input_cut_info', 'log', 'np', 'pathlib', 'prepare_torch_inputs', 'single_input_to_input_cut_info', 'sys', 'to_torch_tensor', 'update_list_or_dict']
+def _build_dynamic_shapes(inputs, input_specs = None):
+    """
+    Build dynamic_shapes for torch.export.export.
+    
+        If input_specs (list of _InputCutInfo from the 'input' parameter) is provided
+        and contains shapes, dimensions marked as -1 (fully dynamic) get Dim.AUTO,
+        dimensions with min/max constraints (e.g. Dimension(1, 10)) get
+        Dim("dI_D", min=..., max=...), and fixed dimensions stay static.
+        When no specs are given returns None so that torch.export.export produces
+        a fully static graph.
+    
+        The input_specs list is flat (one spec per leaf tensor), while inputs may
+        contain nested tuples/lists (e.g. past_key_values). pytree is used to
+        flatten inputs, pair each leaf with its spec, and then unflatten the
+        result back into the original structure that torch.export expects.
+        
+    """
+def _export_torch_model(model, inputs, input_specs = None):
+    """
+    Export a torch.nn.Module using torch.export.export with Dim.AUTO dynamic shapes.
+    """
 def extract_input_info_from_example(args, inputs):
     ...
 def extract_module_extensions(args):