Unexpected output of pooling layer in quantized network CPP export

Required prerequisites

Make sure you've read the documentation. Your issue may be addressed there.
Search the issue tracker and discussions to verify that this hasn't already been reported. +1 or comment there if it has.

What commit version of aidge do you use

aidge_core: dev

Problem description

In a simple CNN, after quantizing the network in int8, I generate a CPP export. I get some compilation warnings (which is not really a problem) that could be fixed. The forward function declares some additional output for each Maxpool layer, here is an example below:

int8_t* _3_MaxPooling2D_1_output_0 = (int8_t*) (mem + _3_MAXPOOLING2D_1_OUTPUT_0_MEM_OFFSET);
int64_t* _3_MaxPooling2D_1_output_1 = (int64_t*) (mem + _3_MAXPOOLING2D_1_OUTPUT_1_MEM_OFFSET);

pooling_forward<_2_PADCONVACT_1_OUTPUT_0_NB_CHANNELS,
                _2_PADCONVACT_1_OUTPUT_0_IN_HEIGHT,
                _2_PADCONVACT_1_OUTPUT_0_IN_WIDTH,
                _3_MAXPOOLING2D_1_OUTPUT_0_NB_OUTPUTS,
                _3_MAXPOOLING2D_1_OUTPUT_0_OUT_HEIGHT,
                _3_MAXPOOLING2D_1_OUTPUT_0_OUT_WIDTH,
                _3_MAXPOOLING2D_1_PADDING_Y,
                _3_MAXPOOLING2D_1_PADDING_X,
                _3_MAXPOOLING2D_1_STRIDE_Y,
                _3_MAXPOOLING2D_1_STRIDE_X,
                _3_MAXPOOLING2D_1_KERNEL_HEIGHT,
                _3_MAXPOOLING2D_1_KERNEL_WIDTH,
                _3_MAXPOOLING2D_1_POOLING_TYPE,
                _3_MAXPOOLING2D_1_ACTIVATION,
                _2_PADCONVACT_1_OUTPUT_0_MEM_CONT_OFFSET,
                _2_PADCONVACT_1_OUTPUT_0_MEM_CONT_SIZE,
                _2_PADCONVACT_1_OUTPUT_0_MEM_WRAP_OFFSET,
                _2_PADCONVACT_1_OUTPUT_0_MEM_WRAP_SIZE,
                _2_PADCONVACT_1_OUTPUT_0_MEM_STRIDE,
                _3_MAXPOOLING2D_1_OUTPUT_0_MEM_CONT_OFFSET,
                _3_MAXPOOLING2D_1_OUTPUT_0_MEM_CONT_SIZE,
                _3_MAXPOOLING2D_1_OUTPUT_0_MEM_WRAP_OFFSET,
                _3_MAXPOOLING2D_1_OUTPUT_0_MEM_WRAP_SIZE,
                _3_MAXPOOLING2D_1_OUTPUT_0_MEM_STRIDE>
                (_2_PadConvAct_1_output_0, _3_MaxPooling2D_1_output_0);

The variable _3_MaxPooling2D_1_output_1 is never used later in the code, and is declared as int64_t which is weird for an int8 quantization. But maybe this uses some memory for nothing.