Fix some operators

aidge_export_arm_cortexm/_Aidge_Arm/kernels/BatchNorm/aidge_batchnorm2d_chw_float32.c

-#include <math.h>
-
-
-void aidge_batchnorm2d_chw_float32 (float* inputs,
-                                float* outputs,
-                                float* input_mean,
-                                float* input_var,
-                                float* scale,
-                                float* bias,
-                                float epsilon,
-                                const int nb_channels,
-                                const int channel_width, const int channel_height)
-{
-    int featureMapSize = channel_width * channel_height;
-    for (int ch = 0; ch < nb_channels; ++ch) 
-    {
-        int ioIndex = ch * featureMapSize;
-        for (int i = ioIndex; i < ioIndex + featureMapSize; i++){
-            outputs[i] = bias[ch];
-        }
-        float var =sqrt(input_var[ch] + epsilon);
-
-        for (int feature = 0; feature<featureMapSize; ++feature) {
-            outputs[ioIndex + feature] += scale[ch] * (inputs[ioIndex + feature]-input_mean[ch]) / var;
-        }
-    
-    }
-}
\ No newline at end of file

aidge_export_arm_cortexm/_Aidge_Arm/kernels/BatchNorm/aidge_batchnorm2d_chw_float32.h

 #include <math.h>
 
-
-void aidge_batchnorm2d_chw_float32 (float* inputs,
-                                float* outputs,
-                                float* input_mean,
-                                float* input_var,
-                                float* scale,
-                                float* bias,
-                                float epsilon,
-                                const int nb_channels,
-                                const int channel_width, const int channel_height)
+void aidge_batchnorm2d_chw_float32 (const float* inputs,
+                                    float* outputs,
+                                    const float* input_mean,
+                                    const float* input_var,
+                                    const float* scale,
+                                    const float* bias,
+                                    float epsilon,
+                                    int nb_channels,
+                                    int channel_width,
+                                    int channel_height)
 {
     int featureMapSize = channel_width * channel_height;
     for (int ch = 0; ch < nb_channels; ++ch) 

aidge_export_arm_cortexm/_Aidge_Arm/kernels/Transform/Gather/aidge_gather_chw_float32.c → aidge_export_arm_cortexm/_Aidge_Arm/kernels/Gather/aidge_gather_chw_float32.h

-
-
-void aidge_gather_chw_float32 (float* inputs, 
-                            float* outputs,
-                            int axis,
-                            int indices[],
-                            int input_dims[],
-                            int size_inputDim,
-                            int indices_size,
-                            unsigned int size)
+void aidge_gather_chw_float32 (const float* inputs,
+                               float* outputs,
+                               int axis,
+                               const int indices[],
+                               const int input_dims[],
+                               int size_inputDim,
+                               int indices_size,
+                               int output_size)
 {
 	axis += (axis >= 0 ) ? 0 : size_inputDim;
 

aidge_export_arm_cortexm/_Aidge_Arm/kernels/Reshape/aidge_reshape.h

 #include <cstring>
-template <unsigned int SIZE, typename Input_T, typename Output_T>
+
+template <unsigned int SIZE, typename T>
 __attribute__((always_inline)) inline static
-void aidge_reshape(Input_T* __restrict input, Output_T* __restrict output, unsigned int size) {
+void aidge_reshape(T* __restrict input, T* __restrict output) {
     // Copy the input data to the output data
-    std::memcpy(output, input, size * sizeof(float));
+    std::memcpy(output, input, SIZE * sizeof(T));
 }
\ No newline at end of file

aidge_export_arm_cortexm/_Aidge_Arm/kernels/Reshape/aidge_reshape_chw_float32.c

-
-void aidge_reshape_chw_float32(float* inputs,
-                           float* outputs,
-                           unsigned int size)
-{
-    for (int i = 0; i < size; i++){
-        outputs[i] = inputs[i];
-    }
-}
\ No newline at end of file

aidge_export_arm_cortexm/_Aidge_Arm/kernels/Sigmoid/aidge_sigmoid_float32.h

-#include <math.h>
-
-
-void aidge_sigmoid_float32 (float* inputs, 
-                            float* outputs,
-                            unsigned int size)
-{
-    for (unsigned int i = 0; i < size; ++i) {
-        outputs[i] = 1 / ( 1 + exp(-inputs[i]) );
-    }
-}
\ No newline at end of file

aidge_export_arm_cortexm/_Aidge_Arm/kernels/Softmax/aidge_softmax_chw_float32.h

 #include <math.h>
 
-void aidge_softmax_chw_float32(const float* inputs, 
-                            float* outputs,
-                            const int inputDims[],
-                            int axis,
-                            const unsigned int size_inputDim,
-                            const unsigned int size)
+void aidge_softmax_chw_float32(const float* inputs,
+                               float* outputs,
+                               const int inputDims[],
+                               int axis,
+                               const unsigned int size_inputDim,
+                               const unsigned int output_size)
 {
     axis += (axis >= 0 ) ? 0 : size_inputDim;
 

aidge_export_arm_cortexm/_Aidge_Arm/kernels/Transform/Transpose/aidge_transpose_chw_float32.c → aidge_export_arm_cortexm/_Aidge_Arm/kernels/Transpose/aidge_transpose_chw_float32.h

-void aidge_transpose_chw_float32   (float* inputs,    
-                                float* outputs,
-                                int input_dims[],
-                                int perm[],
-                                int output_dims[],
-                                unsigned int size_outputDims,
-                                unsigned int size)
+void aidge_transpose_chw_float32 (const float* inputs,
+                                  float* outputs,
+                                  const int input_dims[],
+                                  const int perm[],
+                                  const int output_dims[],
+                                  unsigned int size_outputDims,
+                                  unsigned int output_size)
 {
 	int newStrides[size_outputDims];
 	for (int i = 0; i<size_outputDims;++i){newStrides[i] = 1;}

aidge_export_arm_cortexm/_Aidge_Arm/templates/configuration/batchnorm2d.jinja

@@ -3,12 +3,11 @@
 #define {{ name|upper }}_LAYER_H
 
 {# For layer configuration -#}
-#define {{ name|upper }}_NB_BATCH {{ in_dims[0] }}
-#define {{ name|upper }}_NB_CHANNELS {{ in_dims[1] }}
-#define {{ name|upper }}_CHANNELS_HEIGHT {{ in_dims[2] }}
-#define {{ name|upper }}_CHANNELS_WIDTH {{ in_dims[3] }}
+#define {{ name|upper }}_NB_BATCH {{ in_batch[0] }}
+#define {{ name|upper }}_NB_CHANNELS {{ in_chan[0] }}
+#define {{ name|upper }}_CHANNELS_HEIGHT {{ in_height[0] }}
+#define {{ name|upper }}_CHANNELS_WIDTH {{ in_width[0] }}
 
 #define {{ name|upper }}_EPSILON {{ epsilon }}
 
-
 #endif /* {{ name|upper }}_LAYER_H */
\ No newline at end of file

aidge_export_arm_cortexm/_Aidge_Arm/templates/configuration/gather.jinja

@@ -6,10 +6,11 @@
 
 #define {{name|upper}}_AXIS {{ axis }}
 
-#define {{name|upper}}_INDEXES_DIMS_SIZE {{ indices|length}}
-#define {{name|upper}}_INPUT_DIMS_SIZE {{ input_dims|length}}
-
-#define {{name|upper}}_OUTPUT_SIZE {{ nb_outputs}}
+static const int {{ name|upper }}_INDEXES[] = { {{ indices|join(", ") }} };
+static const int {{ name|upper }}_DIMS[] = { {{ in_dims[0]|join(", ") }} };
 
+#define {{name|upper}}_INDEXES_DIMS_SIZE {{ indices|length }}
+#define {{name|upper}}_INPUT_DIMS_SIZE {{ in_dims[0]|length }}
+#define {{name|upper}}_OUTPUT_SIZE {{ out_size[0] }}
 
 #endif /* {{ name|upper }}_LAYER_H */

aidge_export_arm_cortexm/_Aidge_Arm/templates/configuration/matmul.jinja

@@ -7,9 +7,9 @@
 
 
 {# For layer configuration -#}
-static const int {{ in_name[0]|upper }}_DIMS[] = { {{ in_dims[0]|join(", ") }} };
-static const int {{ in_name[1]|upper}}_DIMS[] = { {{ in_dims[0]|join(", ") }} };
-static const int {{ out_name[0]|upper }}_DIMS[] = { {{ out_dims[0]|join(", ") }} };
+static const int {{name|upper}}_INPUT_A_DIMS[] = { {{ in_dims[0]|join(", ") }} };
+static const int {{name|upper}}_INPUT_B_DIMS[] = { {{ in_dims[1]|join(", ") }} };
+static const int {{name|upper}}_OUTPUT_DIMS[] = { {{ out_dims[0]|join(", ") }} };
 
 #define {{name|upper}}_INPUT_A_DIMS_SIZE {{ in_dims[0]|length}}
 #define {{name|upper}}_INPUT_B_DIMS_SIZE {{ in_dims[1]|length}}

aidge_export_arm_cortexm/_Aidge_Arm/templates/configuration/reshape.jinja

 {#- For name header -#}
 #ifndef {{ name|upper }}_LAYER_H
 #define {{ name|upper }}_LAYER_H
+
 {% include "./_def_io.jinja" %}
 {% include "./_meminfo.jinja" %}
+
 {# For layer configuration -#}
-#define {{ name|upper }}_INPUTS_SIZE {{ nb_inputs }}
-#define {{ name|upper }}_OUTPUTS_SIZE {{ nb_outputs }}
+#define {{ name|upper }}_OUTPUTS_SIZE {{ out_size[0] }}
 
 #endif /* {{ name|upper }}_LAYER_H */
\ No newline at end of file

aidge_export_arm_cortexm/_Aidge_Arm/templates/configuration/sigmoid.jinja

@@ -7,14 +7,6 @@
 {% include "./_meminfo.jinja" %}
 
 {# For layer configuration -#}
-#define {{ name|upper }}_INPUTS_SIZE {{ in_size[0] }}
 #define {{ name|upper }}_OUTPUTS_SIZE {{ out_size[0] }}
 
-
-#define {{in_name[0]|upper}}_NB_DIM {{ in_dims[0]|length}}
-#define {{out_name[0]|upper}}_NB_DIM {{ out_dims[0]|length}}
-
-static const int {{ in_name[0]|upper }}_DIMS[] = { {{ in_dims[0]|join(", ") }} };
-static const int {{ out_name[0]|upper }}_DIMS[] = { {{ out_dims[0]|join(", ") }} };
-
 #endif /* {{ name|upper }}_LAYER_H */

aidge_export_arm_cortexm/_Aidge_Arm/templates/configuration/softmax.jinja

-
-{% include "./_def_io.jinja" %}
-{% include "./_meminfo.jinja" %}
 {#- For name header -#}
 #ifndef {{ name|upper }}_LAYER_H
 #define {{ name|upper }}_LAYER_H
 
+{% include "./_def_io.jinja" %}
+{% include "./_meminfo.jinja" %}
+
 {# For layer configuration -#}
+#define {{ name|upper }}_AXIS {{ axis }}
+
 #define {{ name|upper }}_INPUTS_SIZE {{ in_size[0] }}
 #define {{ name|upper }}_OUTPUTS_SIZE {{ out_size[0] }}
-#define {{ name|upper }}_AXIS {{ axis }}
-#define {{ name|upper }}_INPUT_DIMS_SIZE {{ in_dims[0]|length}}
-static const int {{ name|upper }}_DIMS[] = { {{ in_dims[0] | join(', ') }} };
+
+#define {{ name|upper }}_INPUT_DIMS_SIZE {{ in_dims[0]|length }}
+static const int {{ name|upper }}_INPUT_DIMS[] = { {{ in_dims[0] | join(', ') }} };
+
 #endif /* {{ name|upper }}_LAYER_H */

aidge_export_arm_cortexm/_Aidge_Arm/templates/configuration/transpose.jinja

@@ -3,10 +3,14 @@
 #define {{ name|upper }}_LAYER_H
 
 {# For layer configuration -#}
+static const int {{ name|upper }}_PERMUTATIONS[] = { {{ output_dims_order|join(", ") }} };
 
-#define {{name|upper}}_OUTPUT_DIMS_SIZE {{ output_dims|length}}
+static const int {{ name|upper }}_INPUT_DIMS[] = { {{ in_dims[0]|join(", ") }} };
+static const int {{ name|upper }}_OUTPUT_DIMS[] = { {{ out_dims[0]|join(", ") }} };
 
-#define {{name|upper}}_OUTPUT_SIZE {{ nb_outputs}}
+#define {{name|upper}}_INPUT_DIMS_SIZE {{ in_dims[0]|length }}
+#define {{name|upper}}_OUTPUT_DIMS_SIZE {{ out_dims[0]|length }}
+#define {{name|upper}}_OUTPUT_SIZE {{ out_size[0]}}
 
 
 #endif /* {{ name|upper }}_LAYER_H */

aidge_export_arm_cortexm/_Aidge_Arm/templates/forward_call/batchnorm2d.jinja

-aidge_batchnorm2d_chw_{{dataformat|default("float32") }} 
-({{in_name}},
- {{out_name}}, 
- {{running_mean_name}},
- {{running_var_name}},
- {{weight_name}},
- {{bias_name}}, 
- {{ name|upper }}_EPSILON, 
- {{ name|upper }}_NB_CHANNELS, 
- {{ name|upper }}_CHANNELS_WIDTH, 
- {{ name|upper }}_CHANNELS_HEIGHT);
+{% filter indent(width=4, first=False) %}
+{% include "./_mem_offset.jinja" %}
+aidge_batchnorm2d_chw_float32(
+    {{in_name[0]}},
+    {{out_name[0]}},
+    {{in_name[3]}},
+    {{in_name[4]}},
+    {{in_name[1]}},
+    {{in_name[2]}},
+    {{name|upper}}_EPSILON,
+    {{name|upper}}_NB_CHANNELS,
+    {{name|upper}}_CHANNELS_WIDTH,
+    {{name|upper}}_CHANNELS_HEIGHT);
+{% endfilter %}

aidge_export_arm_cortexm/_Aidge_Arm/templates/forward_call/gather.jinja

-aidge_gather_chw_{{dataformat}} ({{input_name}}, {{output_name}}, {{name|upper}}_AXIS, {{name}}_INDEXES , {{input_name}}_DIMS, {{name|upper}}_INPUT_DIMS_SIZE,{{name|upper}}_INDEXES_DIMS_SIZE,{{name|upper}}_OUTPUT_SIZE);
\ No newline at end of file
+{% filter indent(width=4, first=False) %}
+{% include "./_mem_offset.jinja" %}
+aidge_gather_chw_float32(
+    {{in_name[0]}},
+    {{out_name[0]}},
+    {{name|upper}}_AXIS,
+    {{name|upper}}_INDEXES,
+    {{name|upper}}_DIMS,
+    {{name|upper}}_INPUT_DIMS_SIZE,
+    {{name|upper}}_INDEXES_DIMS_SIZE,
+    {{name|upper}}_OUTPUT_SIZE);
+{% endfilter %}

aidge_export_arm_cortexm/_Aidge_Arm/templates/forward_call/matmul.jinja

 {% filter indent(width=4, first=False) %}
 {% include "./_mem_offset.jinja" %}
-aidge_matmul_chw_{{dataformat| default("float32") }} 
-({{in_name[0]}}, 
-{{in_name[1]}}, 
-{{out_name[0]}}, 
-{{in_name[0]|upper}}_DIMS, 
-{{in_name[1]|upper}}_DIMS, 
-{{out_name[0]|upper}}_DIMS, 
-{{name|upper}}_INPUT_A_DIMS_SIZE, 
-{{name|upper}}_INPUT_B_DIMS_SIZE, 
-{{name|upper}}_OUTPUT_DIMS_SIZE);
+aidge_matmul_chw_float32(
+    {{in_name[0]}},
+    {{in_name[1]}},
+    {{out_name[0]}},
+    {{name|upper}}_INPUT_A_DIMS,
+    {{name|upper}}_INPUT_B_DIMS,
+    {{name|upper}}_OUTPUT_DIMS,
+    {{name|upper}}_INPUT_A_DIMS_SIZE,
+    {{name|upper}}_INPUT_B_DIMS_SIZE,
+    {{name|upper}}_OUTPUT_DIMS_SIZE);
 {% endfilter %}
\ No newline at end of file

aidge_export_arm_cortexm/_Aidge_Arm/templates/forward_call/reshape.jinja

 {% filter indent(width=4, first=False) %}
 {% include "./_mem_offset.jinja" %}
 aidge_reshape<{{name|upper}}_OUTPUTS_SIZE>
-({{input_name}}, {{output_name}}, {{name|upper}}_OUTPUTS_SIZE);
+    ({{in_name[0]}},
+    {{out_name[0]}});
 {% endfilter %}
\ No newline at end of file

aidge_export_arm_cortexm/_Aidge_Arm/templates/forward_call/softmax.jinja

 {% filter indent(width=4, first=False) %}
 {% include "./_mem_offset.jinja" %}
-aidge_softmax_chw_float32({{in_name[0]}}, {{out_name[0]}}, {{name|upper}}_DIMS, {{name|upper}}_AXIS, {{name|upper}}_INPUT_DIMS_SIZE, {{name|upper}}_OUTPUTS_SIZE);
+aidge_softmax_chw_float32(
+    {{in_name[0]}},
+    {{out_name[0]}},
+    {{name|upper}}_INPUT_DIMS,
+    {{name|upper}}_AXIS,
+    {{name|upper}}_INPUT_DIMS_SIZE,
+    {{name|upper}}_OUTPUTS_SIZE);
 {% endfilter %}