[PythonAPI] Adding input and output names options to the wraping function

Those two options can help for pytorch bad consistency in the naming in their ONNX exported layers.
This fix allow the user to provide two list of names for the input and output of the network.

Author: cmoineau

docs/export/CPP_STM32.rst

@@ -1,8 +1,6 @@
 Export: C++/STM32
 =================
 
-**N2D2-IP only: available upon request.**
-
 Export type: ``CPP_STM32``
  C++ export for STM32.
 

docs/quant/pruning.rst

@@ -15,16 +15,16 @@ Example with Python
         :members:
         :inherited-members:
 
-Example of code to use the *PruneCell* in your scripts:
+Example of code to use the :py:class:`n2d2.quantizer.PruneCell` in your scripts:
 
 .. code-block:: python
 
     for cell in model:
-    ### Add Pruning ###
-    if isinstance(cell, n2d2.cells.Conv) or isinstance(cell, n2d2.cells.Fc):
-        cell.quantizer = n2d2.quantizer.PruneCell(prune_mode="Static", threshold=0.3, prune_filler="IterNonStruct")
+        ### Add Pruning ###
+        if isinstance(cell, n2d2.cells.Conv) or isinstance(cell, n2d2.cells.Fc):
+            cell.quantizer = n2d2.quantizer.PruneCell(prune_mode="Static", threshold=0.3, prune_filler="IterNonStruct")
 
-Some explanations with the differents options of the *PruneCell*:
+Some explanations with the differents options of the :py:class:`n2d2.quantizer.PruneCell` :
 
 Pruning mode
 ^^^^^^^^^^^^
@@ -42,7 +42,7 @@ For example, to update each two epochs, write:
 
     n2d2.quantizer.PruneCell(prune_mode="Gradual", threshold=0.3, stepsize=2*DATASET_SIZE)
 
-Where *DATASET_SIZE* is the size of the dataset you are using.
+Where ``DATASET_SIZE`` is the size of the dataset you are using.
 
 Pruning filler
 ^^^^^^^^^^^^^^
@@ -53,7 +53,7 @@ Pruning filler
 - IterNonStruct: all weights below than the ``delta`` factor are pruned. If this is not enough to reach ``threshold``, all the weights below 2 "delta" are pruned and so on...
 
 
-**Important**: With *PruneCell*, ``quant_mode`` and ``range`` are not used.
+**Important**: With :py:class:`n2d2.quantizer.PruneCell`, ``quant_mode`` and ``range`` are not used.
 
 
 Example with INI file

python/pytorch_to_n2d2/pytorch_interface.py

@@ -331,6 +331,8 @@ def __exit__(self, exc_type, exc_value, traceback):
 def wrap(torch_model:torch.nn.Module,
         input_size: Union[list, tuple],
         opset_version:int=11,
+        in_names:list=None,
+        out_names:list=None,
         verbose:bool=False) -> Block:
     """Function generating a ``torch.nn.Module`` which embed a :py:class:`n2d2.cells.DeepNetCell`.
     The torch_model is exported to N2D2 via ONNX.
@@ -341,6 +343,10 @@ def wrap(torch_model:torch.nn.Module,
     :type input_size: ``list``
     :param opset_version: Opset version used to generate the intermediate ONNX file, default=11
     :type opset_version: int, optional
+    :param in_names: Specify specific names for the network inputs
+    :type in_names: list, optional
+    :param out_names: Specify specific names for the network outputs
+    :type in_names: list, optional
     :param verbose: Enable the verbose output of torch onnx export, default=False
     :type verbose: bool, optional
     :return: A custom ``torch.nn.Module`` which embed a :py:class:`n2d2.cells.DeepNetCell`.
@@ -361,6 +367,8 @@ def wrap(torch_model:torch.nn.Module,
             dummy_in,
             raw_model_path,
             verbose=verbose,
+            input_names=in_names,
+            output_names=out_names,
             export_params=True,
             opset_version=opset_version,
             training=torch.onnx.TrainingMode.TRAINING,