Update aidge.export -> aidge_core.export_utils.

aidge_export_arm_cortexm/export.py

@@ -4,7 +4,7 @@ import shutil
 from pathlib import Path
 import numpy as np
 
-from aidge_core.export.code_generation import *
+from aidge_core.export_utils.code_generation import *
 from aidge_export_arm_cortexm.utils import (ROOT, AVAILABLE_BOARDS, has_board, \
                                             OPERATORS_REGISTRY, supported_operators)
 import aidge_export_arm_cortexm.operators
@@ -15,13 +15,13 @@ from aidge_export_arm_cortexm.memory import *
 
 
 
-def export(export_folder_name, 
+def export(export_folder_name,
            graphview,
            scheduler = None,
            board:str ="stm32h7",
            library:str = "aidge",
            mem_wrapping = False):
-    
+
     # Create export directory
     export_folder = Path().absolute() / export_folder_name
     os.makedirs(str(export_folder), exist_ok=True)
@@ -36,7 +36,7 @@ def export(export_folder_name,
         board_path = AVAILABLE_BOARDS[board]
     else:
         raise ValueError(f"{board} not found in the package. Please among those boards: {list(AVAILABLE_BOARDS.keys())}")
-    
+
     # Copy all static files in the export
     shutil.copytree(board_path, str(export_folder), dirs_exist_ok=True)
 

aidge_export_arm_cortexm/operators.py

@@ -8,7 +8,7 @@ from typing import Tuple, List, Union, Dict
 
 import aidge_core
 from aidge_core import ExportNode
-from aidge_core.export.code_generation import *
+from aidge_core.export_utils.code_generation import *
 from aidge_export_arm_cortexm.utils import ROOT, operator_register
 from aidge_export_arm_cortexm.utils.converter import numpy_dtype2ctype, aidge_datatype2dataformat, aidge_datatype2ctype
 from aidge_export_arm_cortexm.utils.generation import *
@@ -78,23 +78,23 @@ class Producer_ARMCortexM:
 
     def export(self, export_file:Path, format:str = "NHWC"):
 
-        if (len(self.values.shape) == 4): 
+        if (len(self.values.shape) == 4):
             # Suppose original dataformat is NCHW
             if format == "NCHW":
-                export_params(self.name, 
-                              self.values.reshape(-1), 
+                export_params(self.name,
+                              self.values.reshape(-1),
                               str(export_file))
             elif format == "NHWC":
-                export_params(self.name, 
-                              np.transpose(self.values, (0, 2, 3, 1)).reshape(-1), 
+                export_params(self.name,
+                              np.transpose(self.values, (0, 2, 3, 1)).reshape(-1),
                               str(export_file))
             else:
                 raise RuntimeError("Producer format export not supported.")
         else:
-            export_params(self.name, 
-                          self.values.reshape(-1), 
+            export_params(self.name,
+                          self.values.reshape(-1),
                           str(export_file))
-            
+
 
 class Scaling():
     class ScalingMode:
@@ -158,7 +158,7 @@ class Scaling():
         assert precision >= 1.0
 
         return power_of_2_divs, precision
-    
+
 
     def __call__(self, mode:str) -> dict:
         """Get dictionnary of scale values in function of the mode
@@ -168,9 +168,9 @@ class Scaling():
         - fixed_point (16 or 32 bits)
         - single_shift
         - double_shift
-        
+
         """
-        
+
         if mode == "floating_point":
             self.scaling = {"scaling_type": "floating_point",
                             "scaling_value": self.scaling_factor}
@@ -189,7 +189,7 @@ class Scaling():
 
             self.scaling = {"scaling_type": "single_shift",
                             "shift_value": shift_value[0]}
-            
+
         elif mode == "double_shift":
             shift_value, _ = self.approximate_shift_scaling(self.scaling_factor, 2)
 
@@ -200,7 +200,7 @@ class Scaling():
             self.scaling = {"scaling_type": "no_scaling"}
 
         return self.scaling
-    
+
 
 @operator_register("ReLU")
 class ReLU_ARMCortexM(ExportNode):
@@ -300,7 +300,7 @@ class Conv_ARMCortexM(ExportNode):
             # Convert the biases to int32
             if self.dataformat != "float32":
                 self.producers[1].values = self.producers[1].values.astype(np.int32)
-            
+
             self.producers[1].export(export_folder / "parameters" / f"{self.producers[1].name}.h")
             list_configs.append(f"parameters/{self.producers[1].name}.h")
 
@@ -323,7 +323,7 @@ class Conv_ARMCortexM(ExportNode):
                 stride=self.stride,
                 padding=self.padding,
                 dilation=self.dilation)
-            
+
         elif self.library == "n2d2":
             # Export configuration file
             generate_file(
@@ -368,9 +368,9 @@ class Conv_ARMCortexM(ExportNode):
                 biases_name=self.inputs[2].name(),
                 outputs_name=self.name
             ))
-            
+
         return list_actions
-   
+
 
 @operator_register("PaddedConv")
 class PaddedConv_ARMCortexM(Conv_ARMCortexM):
@@ -406,7 +406,7 @@ class PaddedConv_ARMCortexM(Conv_ARMCortexM):
         if len(self.outputs_dims[0]) == 4:
             # if dims == [batch, nb_outputs]
             # transform to [nb_outputs, 1, 1]
-            self.outputs_dims[0] = self.outputs_dims[0][1:] 
+            self.outputs_dims[0] = self.outputs_dims[0][1:]
 
 
 @operator_register("ConvReluScaling")
@@ -431,7 +431,7 @@ class ConvReluScaling_ARMCortexM(Conv_ARMCortexM):
         # Impose Single Shift (perhaps change it to have a more modular system)
         self.scaling = Scaling(self.operator.get_attr("scalingFactor"),
                                self.operator.get_attr("quantizedNbBits"))("floating_point")
-    
+
 
 class Pooling_ARMCortexM(ExportNode):
     def __init__(self, node, board, library):
@@ -474,7 +474,7 @@ class Pooling_ARMCortexM(ExportNode):
                 copyfile(str(ROOT / "_Aidge_Arm" / "kernels" / "SupportFunctions" / "aidge_supportfunctions.h"),
                          str(Path(export_folder) / "include"))
 
-            # Export configuration file 
+            # Export configuration file
             generate_file(
                 str(export_folder / "layers" / f"{self.name}.h"),
                 str(ROOT / "_Aidge_Arm" / "templates" / "configuration" / "pooling.jinja"),
@@ -485,8 +485,8 @@ class Pooling_ARMCortexM(ExportNode):
                 stride=self.stride,
                 padding=self.padding,
                 pool_type=self.pool_type)
-            
-            
+
+
         elif self.library == "n2d2":
 
             # Nothing to copy
@@ -572,7 +572,7 @@ class FC_ARMCortexM(ExportNode):
             # if dims == [batch, nb_channels, height, width]
             # transform to [nb_channels, height, width]
             self.inputs_dims[0] = self.inputs_dims[0][1:]
-            
+
             # It also means that we need to change the dataformat of the weights
             weights = self.producers[0].values
             if len(weights.shape) == 2:
@@ -623,7 +623,7 @@ class FC_ARMCortexM(ExportNode):
                 channel_height=self.inputs_dims[0][1],
                 channel_width=self.inputs_dims[0][2],
                 nb_outputs=self.outputs_dims[0][0])
-            
+
         elif self.library == "n2d2":
 
             # Export configuration file
@@ -635,7 +635,7 @@ class FC_ARMCortexM(ExportNode):
                 output_dims=self.outputs_dims[0],
                 activation=self.activation,
                 **self.scaling)
-        
+
         return list_configs
 
     def forward(self, list_actions:list):
@@ -666,7 +666,7 @@ class FC_ARMCortexM(ExportNode):
 
 
         return list_actions
-    
+
 
 @operator_register("FcScaling")
 class FCScaling_ARMCortexM(FC_ARMCortexM):