diff --git a/include/aidge/backend/cpu/operator/AvgPoolingImpl_kernels.hpp b/include/aidge/backend/cpu/operator/AvgPoolingImpl_kernels.hpp
index 0d73cb912a0b8218c29d1f533b674a8ea5005d26..e7bc3a2b845d077877684f75d7980fcd1958eb6e 100644
--- a/include/aidge/backend/cpu/operator/AvgPoolingImpl_kernels.hpp
+++ b/include/aidge/backend/cpu/operator/AvgPoolingImpl_kernels.hpp
@@ -79,8 +79,8 @@ void AvgPoolingImpl2D_cpu_forward_kernel(const std::array<DimSize_t, 2>& strideD
#ifdef _OPENMP
#pragma omp parallel for collapse(2) if (dims[0] * dims[1] > 32)
#endif
- for (std::size_t batch = 0; batch < dims[0]; ++batch) {
- for (std::size_t ch = 0; ch < dims[1]; ++ch) {
+ for (int batch = 0; batch < static_cast<int>(dims[0]); ++batch) {
+ for (int ch = 0; ch < static_cast<int>(dims[1]); ++ch) {
const std::size_t oIndex = (ch + batch * dims[1]) * oxSize * oySize;
const std::size_t iIndex = (ch + batch * dims[1]) * dims[2] * dims[3];
diff --git a/include/aidge/backend/cpu/operator/BatchNormImpl_kernels.hpp b/include/aidge/backend/cpu/operator/BatchNormImpl_kernels.hpp
index 7bb7971e35916b5ecf9c59ad3fe55965f53eed91..105a33007bae830507128804641b8feb16bc0848 100644
--- a/include/aidge/backend/cpu/operator/BatchNormImpl_kernels.hpp
+++ b/include/aidge/backend/cpu/operator/BatchNormImpl_kernels.hpp
@@ -56,8 +56,8 @@ void BatchNormImpl2D_cpu_forward_kernel(float epsilon, float momentum, const std
#ifdef _OPENMP
#pragma omp parallel for collapse(2) if (nbBatch * nbChannels > 32)
#endif
- for (std::size_t batch = 0; batch < nbBatch; ++batch) {
- for (std::size_t ch = 0; ch < nbChannels; ++ch) {
+ for (int batch = 0; batch < static_cast<int>(nbBatch); ++batch) {
+ for (int ch = 0; ch < static_cast<int>(nbChannels); ++ch) {
const std::size_t ioIndex = (ch + batch*nbChannels) * featureMapSize;
std::fill(output + ioIndex, output + ioIndex + featureMapSize, shift[ch]);
const P var = std::sqrt(batchVar[ch] + static_cast<P>(epsilon));
diff --git a/include/aidge/backend/cpu/operator/ConvDepthWiseImpl_kernels.hpp b/include/aidge/backend/cpu/operator/ConvDepthWiseImpl_kernels.hpp
index b16a819b4b42127dddda0659099018a494d06bc9..3019b1d254f4102aabbcfcf4b59b40f10732a18b 100644
--- a/include/aidge/backend/cpu/operator/ConvDepthWiseImpl_kernels.hpp
+++ b/include/aidge/backend/cpu/operator/ConvDepthWiseImpl_kernels.hpp
@@ -68,8 +68,8 @@ void ConvDepthWiseImpl1D_cpu_forward_kernel(const std::array<DimSize_t, 1>& stri
#ifdef _OPENMP
#pragma omp parallel for collapse(2) if (inputDims[0] * inputDims[1] > 32)
#endif
- for (std::size_t batch = 0; batch < inputDims[0]; ++batch) {
- for (std::size_t ch = 0; ch < inputDims[1]; ++ch) {
+ for (int batch = 0; batch < static_cast<int>(inputDims[0]); ++batch) {
+ for (int ch = 0; ch < static_cast<int>(inputDims[1]); ++ch) {
const std::size_t oIndex = (ch + batch*inputDims[1]) * oxSize;
B biasVal = (biases != nullptr) ? biases[ch] : B(0);
std::fill(output + oIndex, output+(oIndex+oxSize), biasVal);
@@ -158,8 +158,8 @@ void ConvDepthWiseImpl2D_cpu_forward_kernel(const std::array<DimSize_t, 2>& stri
#ifdef _OPENMP
#pragma omp parallel for collapse(2) if (inputDims[0] * inputDims[1] > 32)
#endif
- for (std::size_t batch = 0; batch < inputDims[0]; ++batch) {
- for (std::size_t ch = 0; ch < inputDims[1]; ++ch) {
+ for (int batch = 0; batch < static_cast<int>(inputDims[0]); ++batch) {
+ for (int ch = 0; ch < static_cast<int>(inputDims[1]); ++ch) {
B biasVal = (biases != nullptr) ? biases[ch] : B(0);
std::size_t oIndex = (ch + batch*inputDims[1]) * outChannels_s;
@@ -201,8 +201,8 @@ void ConvDepthWiseImpl2D_cpu_forward_kernel(const std::array<DimSize_t, 2>& stri
#ifdef _OPENMP
#pragma omp parallel for collapse(2) if (inputDims[0] * inputDims[1] > 32)
#endif
- for (std::size_t batch = 0; batch < inputDims[0]; ++batch) {
- for (std::size_t ch = 0; ch < inputDims[1]; ++ch) {
+ for (int batch = 0; batch < static_cast<int>(inputDims[0]); ++batch) {
+ for (int ch = 0; ch < static_cast<int>(inputDims[1]); ++ch) {
B biasVal = (biases != nullptr) ? biases[ch] : B(0);
std::size_t oIndex = (ch + batch*inputDims[1]) * outChannels_s;
@@ -226,8 +226,8 @@ void ConvDepthWiseImpl2D_cpu_forward_kernel(const std::array<DimSize_t, 2>& stri
#ifdef _OPENMP
#pragma omp parallel for collapse(2) if (inputDims[0] * inputDims[1] > 32)
#endif
- for (std::size_t batch = 0; batch < inputDims[0]; ++batch) {
- for (std::size_t ch = 0; ch < inputDims[1]; ++ch) {
+ for (int batch = 0; batch < static_cast<int>(inputDims[0]); ++batch) {
+ for (int ch = 0; ch < static_cast<int>(inputDims[1]); ++ch) {
const std::size_t oIndex = (ch + batch*inputDims[1]) * outChannels_s;
const std::size_t iIndex = (ch + batch*inputDims[1]) * inputDims[2] * inputDims[3];
const std::size_t wIndex = ch * kernelDims[0] * kernelDims[1];
diff --git a/include/aidge/backend/cpu/operator/ConvImpl_kernels.hpp b/include/aidge/backend/cpu/operator/ConvImpl_kernels.hpp
index b1cd006eae2d0ed9a224b8bdbf5b267472795720..d72761600f261e1222c287e051af98561819011e 100644
--- a/include/aidge/backend/cpu/operator/ConvImpl_kernels.hpp
+++ b/include/aidge/backend/cpu/operator/ConvImpl_kernels.hpp
@@ -62,8 +62,8 @@ void ConvImpl1D_cpu_forward_kernel(const array<DimSize_t, 1> &strideDim,
#ifdef _OPENMP
#pragma omp parallel for collapse(2) if (inputDims[0] * outChannels > 32)
#endif
- for (std::size_t batch = 0; batch < inputDims[0]; ++batch) {
- for (std::size_t outCh = 0; outCh < outChannels; ++outCh) {
+ for (int batch = 0; batch < static_cast<int>(inputDims[0]); ++batch) {
+ for (int outCh = 0; outCh < static_cast<int>(outChannels); ++outCh) {
const std::size_t oIndex = (outCh + batch * outChannels) * oxSize;
// If bias = nullptr, set B(0)
B biasVal = (biases != nullptr) ? biases[outCh] : B(0);
@@ -484,8 +484,8 @@ void ConvImpl2D_cpu_forward_kernel(const array<DimSize_t, 2> &strideDims,
#ifdef _OPENMP
#pragma omp parallel for collapse(2) if (inputDims[0] * outChannels > 32)
#endif
- for (std::size_t batch = 0; batch < inputDims[0]; ++batch) {
- for (std::size_t outCh = 0; outCh < outChannels; ++outCh) {
+ for (int batch = 0; batch < static_cast<int>(inputDims[0]); ++batch) {
+ for (int outCh = 0; outCh < static_cast<int>(outChannels); ++outCh) {
std::size_t oIndex = (outCh + batch*inputDims[1]) * outChannels_s;
// If bias = nullptr, set B(0)
@@ -573,8 +573,8 @@ void ConvImpl2D_cpu_forward_kernel(const array<DimSize_t, 2> &strideDims,
#ifdef _OPENMP
#pragma omp parallel for collapse(2) if (inputDims[0] * outChannels > 32)
#endif
- for (std::size_t batch = 0; batch < inputDims[0]; ++batch) {
- for (std::size_t outCh = 0; outCh < outChannels; ++outCh) {
+ for (int batch = 0; batch < static_cast<int>(inputDims[0]); ++batch) {
+ for (int outCh = 0; outCh < static_cast<int>(outChannels); ++outCh) {
std::size_t oIndex = (outCh + batch*inputDims[1]) * outChannels_s;
// If bias = nullptr, set B(0)
@@ -609,8 +609,8 @@ void ConvImpl2D_cpu_forward_kernel(const array<DimSize_t, 2> &strideDims,
#ifdef _OPENMP
#pragma omp parallel for collapse(2) if (inputDims[0] * outChannels > 32)
#endif
- for (std::size_t batch = 0; batch < inputDims[0]; ++batch) {
- for (std::size_t outCh = 0; outCh < outChannels; ++outCh) {
+ for (int batch = 0; batch < static_cast<int>(inputDims[0]); ++batch) {
+ for (int outCh = 0; outCh < static_cast<int>(outChannels); ++outCh) {
std::size_t oIndex = (outCh + batch*inputDims[1]) * outChannels_s;
// If bias = nullptr, set B(0)
diff --git a/include/aidge/backend/cpu/operator/GlobalAveragePoolingImpl_kernels.hpp b/include/aidge/backend/cpu/operator/GlobalAveragePoolingImpl_kernels.hpp
index 3915adb3a0fcfd8cef0bf78761b2169272e1c211..8ff1ad08b0897dbc3d89d67632a555effba73a85 100644
--- a/include/aidge/backend/cpu/operator/GlobalAveragePoolingImpl_kernels.hpp
+++ b/include/aidge/backend/cpu/operator/GlobalAveragePoolingImpl_kernels.hpp
@@ -77,8 +77,8 @@ void GlobalAveragePoolingImpl_cpu_forward_kernel(const std::shared_ptr<Tensor>&
#ifdef _OPENMP
#pragma omp parallel for collapse(2) if (dims[0] * dims[1] > 32)
#endif
- for (DimSize_t batch = 0; batch < dims[0]; ++batch) {
- for (DimSize_t channel = 0; channel < dims[1]; ++channel) {
+ for (int batch = 0; batch < static_cast<int>(dims[0]); ++batch) {
+ for (int channel = 0; channel < static_cast<int>(dims[1]); ++channel) {
const I *filter_start = std::next(
input, (batch * in_batch_nb_elems) + (channel * in_channel_nb_elems));
output[batch * out_batch_nb_elems + channel] = castFromFloat<O>(stableMean<I>(filter_start, in_channel_nb_elems));
diff --git a/include/aidge/backend/cpu/operator/MaxPoolingImpl_kernels.hpp b/include/aidge/backend/cpu/operator/MaxPoolingImpl_kernels.hpp
index 9772b0abd2dfd0e5a6b2e11a856f734592e478db..b5f219f9086f387f86769a582c2a2cd6aaa42d9f 100644
--- a/include/aidge/backend/cpu/operator/MaxPoolingImpl_kernels.hpp
+++ b/include/aidge/backend/cpu/operator/MaxPoolingImpl_kernels.hpp
@@ -69,8 +69,8 @@ void MaxPoolingImpl2D_cpu_forward_kernel(
#ifdef _OPENMP
#pragma omp parallel for collapse(2) if (dims[0] * dims[1] > 32)
#endif
- for (std::size_t batch = 0; batch < dims[0]; ++batch){
- for (std::size_t channel = 0; channel < dims[1]; ++channel){
+ for (int batch = 0; batch < static_cast<int>(dims[0]); ++batch){
+ for (int channel = 0; channel < static_cast<int>(dims[1]); ++channel){
auto batchChannelIndex = (channel + batch * dims[1]);
const std::size_t outputBaseIndex = batchChannelIndex * outXSize * outYSize;
const std::size_t inputBaseIndex = batchChannelIndex * dims[2] * dims[3];
diff --git a/include/aidge/backend/cpu/operator/SoftmaxImpl_kernels.hpp b/include/aidge/backend/cpu/operator/SoftmaxImpl_kernels.hpp
index e74f3518bf0f394d17d89542b2a2221047beb0af..ab6790e257b04a2fb1ee3d3ed57c5c7220c6c456 100644
--- a/include/aidge/backend/cpu/operator/SoftmaxImpl_kernels.hpp
+++ b/include/aidge/backend/cpu/operator/SoftmaxImpl_kernels.hpp
@@ -40,8 +40,8 @@ void SoftmaxImpl_cpu_forward_kernel(std::size_t axisIdx, const std::vector<DimSi
#ifdef _OPENMP
#pragma omp parallel for collapse(2) if (preAxisElems * postAxisElems > 32)
#endif
- for (std::size_t i = 0; i < preAxisElems; ++i) {
- for (std::size_t j = 0; j < postAxisElems; ++j) {
+ for (int i = 0; i < static_cast<int>(preAxisElems); ++i) {
+ for (int j = 0; j < static_cast<int>(postAxisElems); ++j) {
I maxVal = input[i * inputDims[axisIdx] * postAxisElems + j];
for (std::size_t k = 1; k < inputDims[axisIdx]; ++k) {
std::size_t inIdx = i * inputDims[axisIdx] * postAxisElems + k * postAxisElems + j;