From 72f494313f6948b1f51f22dd951eb7fa4a96afa4 Mon Sep 17 00:00:00 2001
From: NAUD Maxence <maxence.naud@cea.fr>
Date: Tue, 5 Nov 2024 14:45:28 +0000
Subject: [PATCH] [add] int32 and int64 types to const_filler and
uniform_filler
---
python_binding/filler/pybind_Filler.cpp | 50 ++++++++++++++++++-------
src/filler/ConstantFiller.cpp | 3 +-
src/filler/UniformFiller.cpp | 17 +++++++--
3 files changed, 52 insertions(+), 18 deletions(-)
diff --git a/python_binding/filler/pybind_Filler.cpp b/python_binding/filler/pybind_Filler.cpp
index a85c0d6cd..dbf9a4845 100644
--- a/python_binding/filler/pybind_Filler.cpp
+++ b/python_binding/filler/pybind_Filler.cpp
@@ -30,11 +30,17 @@ void init_Filler(py::module &m) {
[](std::shared_ptr<Tensor> tensor, py::object value) -> void {
switch (tensor->dataType()) {
case DataType::Float64:
- constantFiller<double>(tensor, value.cast<double>());
+ constantFiller<cpptype_t<DataType::Float64>>(tensor, value.cast<cpptype_t<DataType::Float64>>());
break;
case DataType::Float32:
- constantFiller<float>(tensor, value.cast<float>());
+ constantFiller<cpptype_t<DataType::Float32>>(tensor, value.cast<cpptype_t<DataType::Float32>>());
break;
+ case DataType::Int64:
+ constantFiller<cpptype_t<DataType::Int64>>(tensor, value.cast<cpptype_t<DataType::Int64>>());
+ break;
+ case DataType::Int32:
+ constantFiller<cpptype_t<DataType::Int32>>(tensor, value.cast<cpptype_t<DataType::Int32>>());
+ break;
default:
AIDGE_THROW_OR_ABORT(
py::value_error,
@@ -44,14 +50,14 @@ void init_Filler(py::module &m) {
py::arg("tensor"), py::arg("value"))
.def(
"normal_filler",
- [](std::shared_ptr<Tensor> tensor, double mean,
- double stdDev) -> void {
+ [](std::shared_ptr<Tensor> tensor, py::object mean,
+ py::object stdDev) -> void {
switch (tensor->dataType()) {
case DataType::Float64:
- normalFiller<double>(tensor, mean, stdDev);
+ normalFiller<cpptype_t<DataType::Float64>>(tensor, mean.cast<cpptype_t<DataType::Float64>>(), stdDev.cast<cpptype_t<DataType::Float64>>());
break;
case DataType::Float32:
- normalFiller<float>(tensor, mean, stdDev);
+ normalFiller<cpptype_t<DataType::Float64>>(tensor, mean.cast<cpptype_t<DataType::Float32>>(), stdDev.cast<cpptype_t<DataType::Float32>>());
break;
default:
AIDGE_THROW_OR_ABORT(
@@ -60,23 +66,39 @@ void init_Filler(py::module &m) {
}
},
py::arg("tensor"), py::arg("mean") = 0.0, py::arg("stdDev") = 1.0)
- .def(
- "uniform_filler",
- [](std::shared_ptr<Tensor> tensor, double min, double max) -> void {
+ .def("uniform_filler", [] (std::shared_ptr<Tensor> tensor, py::object min, py::object max) -> void {
+ if (py::isinstance<py::int_>(min) && py::isinstance<py::int_>(max)) {
switch (tensor->dataType()) {
- case DataType::Float64:
- uniformFiller<double>(tensor, min, max);
+ case DataType::Int32:
+ uniformFiller<std::int32_t>(tensor, min.cast<std::int32_t>(), max.cast<std::int32_t>());
+ break;
+ case DataType::Int64:
+ uniformFiller<std::int64_t>(tensor, min.cast<std::int64_t>(), max.cast<std::int64_t>());
+ break;
+ default:
+ AIDGE_THROW_OR_ABORT(
+ py::value_error,
+ "Data type is not supported for Uniform filler.");
break;
+ }
+ } else if (py::isinstance<py::float_>(min) && py::isinstance<py::float_>(max)) {
+ switch (tensor->dataType()) {
case DataType::Float32:
- uniformFiller<float>(tensor, min, max);
+ uniformFiller<float>(tensor, min.cast<float>(), max.cast<float>());
+ break;
+ case DataType::Float64:
+ uniformFiller<double>(tensor, min.cast<double>(), max.cast<double>());
break;
default:
AIDGE_THROW_OR_ABORT(
py::value_error,
"Data type is not supported for Uniform filler.");
+ break;
}
- },
- py::arg("tensor"), py::arg("min"), py::arg("max"))
+ } else {
+ AIDGE_THROW_OR_ABORT(py::value_error,"Input must be either an int or a float.");
+ }
+ }, py::arg("tensor"), py::arg("min"), py::arg("max"))
.def(
"xavier_uniform_filler",
[](std::shared_ptr<Tensor> tensor, py::object scaling,
diff --git a/src/filler/ConstantFiller.cpp b/src/filler/ConstantFiller.cpp
index 1e992f4a1..b2118866f 100644
--- a/src/filler/ConstantFiller.cpp
+++ b/src/filler/ConstantFiller.cpp
@@ -39,6 +39,7 @@ void Aidge::constantFiller(std::shared_ptr<Aidge::Tensor> tensor, T constantValu
tensor->copyCastFrom(tensorWithValues);
}
-
+template void Aidge::constantFiller<std::int32_t>(std::shared_ptr<Aidge::Tensor>, std::int32_t);
+template void Aidge::constantFiller<std::int64_t>(std::shared_ptr<Aidge::Tensor>, std::int64_t);
template void Aidge::constantFiller<float>(std::shared_ptr<Aidge::Tensor>, float);
template void Aidge::constantFiller<double>(std::shared_ptr<Aidge::Tensor>, double);
diff --git a/src/filler/UniformFiller.cpp b/src/filler/UniformFiller.cpp
index a942f59d7..1951fcc62 100644
--- a/src/filler/UniformFiller.cpp
+++ b/src/filler/UniformFiller.cpp
@@ -8,8 +8,9 @@
* SPDX-License-Identifier: EPL-2.0
*
********************************************************************************/
+#include <cstdint> // std::int32_t
#include <memory>
-#include <random> // normal_distribution, uniform_real_distribution
+#include <random> // normal_distribution, uniform_real_distribution
#include "aidge/data/Tensor.hpp"
#include "aidge/filler/Filler.hpp"
@@ -19,10 +20,16 @@ template <typename T>
void Aidge::uniformFiller(std::shared_ptr<Aidge::Tensor> tensor, T min, T max) {
AIDGE_ASSERT(tensor->getImpl(),
"Tensor got no implementation, cannot fill it.");
- AIDGE_ASSERT(NativeType<T>::type == tensor->dataType(), "Wrong data type");
+ AIDGE_ASSERT(NativeType<T>::type == tensor->dataType(), "Wrong data type {} and {}",NativeType<T>::type, tensor->dataType());
- std::uniform_real_distribution<T> uniformDist(min, max);
+ using DistType = typename std::conditional<
+ std::is_integral<T>::value,
+ std::uniform_int_distribution<T>,
+ std::uniform_real_distribution<T>
+ >::type;
+
+ DistType uniformDist(min, max);
std::shared_ptr<Aidge::Tensor> cpyTensor;
// Create cpy only if tensor not on CPU
@@ -42,3 +49,7 @@ template void Aidge::uniformFiller<float>(std::shared_ptr<Aidge::Tensor>, float,
float);
template void Aidge::uniformFiller<double>(std::shared_ptr<Aidge::Tensor>,
double, double);
+template void Aidge::uniformFiller<std::int32_t>(std::shared_ptr<Aidge::Tensor>,
+ std::int32_t, std::int32_t);
+template void Aidge::uniformFiller<std::int64_t>(std::shared_ptr<Aidge::Tensor>,
+ std::int64_t, std::int64_t);
\ No newline at end of file
--
GitLab