From b3667685c0c56c35f113a5117f0b03bb616d5a0a Mon Sep 17 00:00:00 2001
From: cmoineau <cyril.moineau@cea.fr>
Date: Thu, 21 Mar 2024 16:00:53 +0000
Subject: [PATCH] Add a Random Generator handler class and use it for fillers.
---
include/aidge/filler/Filler.hpp | 2 --
include/aidge/utils/Random.hpp | 50 ++++++++++++++++++++------
python_binding/pybind_core.cpp | 3 ++
python_binding/utils/pybind_Random.cpp | 24 +++++++++++++
src/data/DataProvider.cpp | 8 ++---
src/filler/HeFiller.cpp | 6 ++--
src/filler/NormalFiller.cpp | 5 ++-
src/filler/UniformFiller.cpp | 6 ++--
src/filler/XavierFiller.cpp | 20 +++++------
src/utils/Random.cpp | 22 ++++++++++++
10 files changed, 108 insertions(+), 38 deletions(-)
create mode 100644 python_binding/utils/pybind_Random.cpp
create mode 100644 src/utils/Random.cpp
diff --git a/include/aidge/filler/Filler.hpp b/include/aidge/filler/Filler.hpp
index c7b12a35c..51d01d87f 100644
--- a/include/aidge/filler/Filler.hpp
+++ b/include/aidge/filler/Filler.hpp
@@ -41,12 +41,10 @@ enum VarianceNorm { FanIn, Average, FanOut };
template <typename T>
void constantFiller(std::shared_ptr<Tensor> tensor, T constantValue);
-// TODO: Keep template or use switch case depending on Tensor datatype ?
template <typename T>
void normalFiller(std::shared_ptr<Tensor> tensor, double mean = 0.0,
double stdDev = 1.0);
-// TODO: Keep template or use switch case depending on Tensor datatype ?
template <typename T>
void uniformFiller(std::shared_ptr<Tensor> tensor, T min, T max);
diff --git a/include/aidge/utils/Random.hpp b/include/aidge/utils/Random.hpp
index 704609c0c..73cbd1453 100644
--- a/include/aidge/utils/Random.hpp
+++ b/include/aidge/utils/Random.hpp
@@ -9,23 +9,53 @@
*
********************************************************************************/
-
#ifndef AIDGE_RANDOM_H_
#define AIDGE_RANDOM_H_
-
#include <algorithm>
-#include <vector>
#include <random>
+#include <vector>
+namespace Aidge {
namespace Random {
- void randShuffle(std::vector<unsigned int>& vec) {
- std::random_device rd;
- std::mt19937 g(rd());
- std::shuffle(vec.begin(), vec.end(), g);
- }
-
+/**
+ * @brief Generator is a class created to handle only one Mersenne Twister
+ * pseudo-random number generator for the whole Aidge framework.
+ *
+ * All of its method are static. You can set a random seed and access the
+ * generator.
+ * By default, the random seed is set to 0 but selected randomly.
+ *
+ */
+class Generator {
+ public:
+ /**
+ * @brief Set a seed to the pseudo-random number generator.
+ *
+ * @return std::mt19937&
+ */
+ static void setSeed(unsigned int seed);
+ static unsigned int getSeed() { return seed; };
+ /**
+ * @brief Return a Mersenne Twister pseudo-random number generator.
+ * You can set the seed of this generator using ``setSeed`` method.
+ *
+ * @return std::mt19937&
+ */
+ static std::mt19937& get() { return generator; };
+
+ private:
+ // Mersenne Twister pseudo-random number generator
+ static std::mt19937 generator;
+ static unsigned int seed;
+};
+
+inline void randShuffle(std::vector<unsigned int>& vec) {
+ std::shuffle(vec.begin(), vec.end(), Aidge::Random::Generator::get());
}
-#endif //AIDGE_RANDOM_H_
\ No newline at end of file
+} // namespace Random
+} // namespace Aidge
+
+#endif // AIDGE_RANDOM_H_
diff --git a/python_binding/pybind_core.cpp b/python_binding/pybind_core.cpp
index 80f0cc44f..5ffa8f6b4 100644
--- a/python_binding/pybind_core.cpp
+++ b/python_binding/pybind_core.cpp
@@ -16,6 +16,7 @@
namespace py = pybind11;
namespace Aidge {
+void init_Random(py::module&);
void init_Data(py::module&);
void init_Database(py::module&);
void init_DataProvider(py::module&);
@@ -73,6 +74,8 @@ void init_TensorUtils(py::module&);
void init_Filler(py::module&);
void init_Aidge(py::module& m) {
+ init_Random(m);
+
init_Data(m);
init_Database(m);
init_DataProvider(m);
diff --git a/python_binding/utils/pybind_Random.cpp b/python_binding/utils/pybind_Random.cpp
new file mode 100644
index 000000000..a1956d2d1
--- /dev/null
+++ b/python_binding/utils/pybind_Random.cpp
@@ -0,0 +1,24 @@
+/********************************************************************************
+ * Copyright (c) 2023 CEA-List
+ *
+ * This program and the accompanying materials are made available under the
+ * terms of the Eclipse Public License 2.0 which is available at
+ * http://www.eclipse.org/legal/epl-2.0.
+ *
+ * SPDX-License-Identifier: EPL-2.0
+ *
+ ********************************************************************************/
+
+#include <pybind11/pybind11.h>
+#include "aidge/utils/Random.hpp"
+
+namespace py = pybind11;
+
+namespace Aidge {
+
+void init_Random(py::module &m) {
+ auto mRand = m.def_submodule("random", "Random module.");
+ py::class_<Random::Generator>(mRand, "Generator")
+ .def_static("set_seed", Random::Generator::setSeed);
+}
+} // namespace Aidge
diff --git a/src/data/DataProvider.cpp b/src/data/DataProvider.cpp
index 7783ed86c..5c3d1d7ef 100644
--- a/src/data/DataProvider.cpp
+++ b/src/data/DataProvider.cpp
@@ -41,8 +41,8 @@ Aidge::DataProvider::DataProvider(const Aidge::Database& database, const std::si
}
// Compute the number of bacthes depending on mDropLast boolean
- mNbBatch = (mDropLast) ?
- static_cast<std::size_t>(std::floor(mNbItems / mBatchSize)) :
+ mNbBatch = (mDropLast) ?
+ static_cast<std::size_t>(std::floor(mNbItems / mBatchSize)) :
static_cast<std::size_t>(std::ceil(mNbItems / mBatchSize));
}
@@ -98,7 +98,7 @@ std::vector<std::shared_ptr<Aidge::Tensor>> Aidge::DataProvider::readBatch() con
void Aidge::DataProvider::setBatches(){
-
+
mBatches.clear();
mBatches.resize(mNbItems);
std::iota(mBatches.begin(),
@@ -106,7 +106,7 @@ void Aidge::DataProvider::setBatches(){
0U);
if (mShuffle){
- Random::randShuffle(mBatches);
+ Aidge::Random::randShuffle(mBatches);
}
if (mNbItems % mBatchSize !=0){ // The last batch is not full
diff --git a/src/filler/HeFiller.cpp b/src/filler/HeFiller.cpp
index e49386b49..74d681f1a 100644
--- a/src/filler/HeFiller.cpp
+++ b/src/filler/HeFiller.cpp
@@ -13,6 +13,7 @@
#include "aidge/data/Tensor.hpp"
#include "aidge/filler/Filler.hpp"
+#include "aidge/utils/Random.hpp"
template <typename T>
void Aidge::heFiller(std::shared_ptr<Aidge::Tensor> tensor,
@@ -36,9 +37,6 @@ void Aidge::heFiller(std::shared_ptr<Aidge::Tensor> tensor,
? meanNorm / ((fanIn + fanOut) / 2.0)
: meanNorm / fanOut);
- std::random_device rd;
- std::mt19937 gen(rd()); // Mersenne Twister pseudo-random number generator
-
std::normal_distribution<T> normalDist(mean, stdDev);
std::shared_ptr<Tensor> cpyTensor;
@@ -48,7 +46,7 @@ void Aidge::heFiller(std::shared_ptr<Aidge::Tensor> tensor,
// Setting values
for (std::size_t idx = 0; idx < tensorWithValues.size(); ++idx) {
- tensorWithValues.set<T>(idx, scaling*normalDist(gen));
+ tensorWithValues.set<T>(idx, scaling*normalDist(Aidge::Random::Generator::get()));
}
// Copy values back to the original tensors (actual copy only if needed)
diff --git a/src/filler/NormalFiller.cpp b/src/filler/NormalFiller.cpp
index 0fadbd134..f30b32431 100644
--- a/src/filler/NormalFiller.cpp
+++ b/src/filler/NormalFiller.cpp
@@ -13,6 +13,7 @@
#include "aidge/data/Tensor.hpp"
#include "aidge/filler/Filler.hpp"
+#include "aidge/utils/Random.hpp"
template <typename T>
void Aidge::normalFiller(std::shared_ptr<Aidge::Tensor> tensor, double mean,
@@ -20,8 +21,6 @@ void Aidge::normalFiller(std::shared_ptr<Aidge::Tensor> tensor, double mean,
AIDGE_ASSERT(tensor->getImpl(),
"Tensor got no implementation, cannot fill it.");
AIDGE_ASSERT(NativeType<T>::type == tensor->dataType(), "Wrong data type");
- std::random_device rd;
- std::mt19937 gen(rd()); // Mersenne Twister pseudo-random number generator
std::normal_distribution<T> normalDist(mean, stdDev);
@@ -32,7 +31,7 @@ void Aidge::normalFiller(std::shared_ptr<Aidge::Tensor> tensor, double mean,
// Setting values
for (std::size_t idx = 0; idx < tensorWithValues.size(); ++idx) {
- tensorWithValues.set<T>(idx, normalDist(gen));
+ tensorWithValues.set<T>(idx, normalDist(Aidge::Random::Generator::get()));
}
// Copy values back to the original tensors (actual copy only if needed)
diff --git a/src/filler/UniformFiller.cpp b/src/filler/UniformFiller.cpp
index e45d6f13e..a942f59d7 100644
--- a/src/filler/UniformFiller.cpp
+++ b/src/filler/UniformFiller.cpp
@@ -13,14 +13,14 @@
#include "aidge/data/Tensor.hpp"
#include "aidge/filler/Filler.hpp"
+#include "aidge/utils/Random.hpp"
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");
- std::random_device rd;
- std::mt19937 gen(rd()); // Mersenne Twister pseudo-random number generator
+
std::uniform_real_distribution<T> uniformDist(min, max);
@@ -31,7 +31,7 @@ void Aidge::uniformFiller(std::shared_ptr<Aidge::Tensor> tensor, T min, T max) {
// Setting values
for (std::size_t idx = 0; idx < tensorWithValues.size(); ++idx) {
- tensorWithValues.set<T>(idx, uniformDist(gen));
+ tensorWithValues.set<T>(idx, uniformDist(Aidge::Random::Generator::get()));
}
// Copy values back to the original tensors (actual copy only if needed)
diff --git a/src/filler/XavierFiller.cpp b/src/filler/XavierFiller.cpp
index f1c5d17e8..a1de15971 100644
--- a/src/filler/XavierFiller.cpp
+++ b/src/filler/XavierFiller.cpp
@@ -13,10 +13,11 @@
#include "aidge/data/Tensor.hpp"
#include "aidge/filler/Filler.hpp"
+#include "aidge/utils/Random.hpp"
template <typename T>
-void Aidge::xavierUniformFiller(std::shared_ptr<Aidge::Tensor> tensor, T scaling,
- Aidge::VarianceNorm varianceNorm) {
+void Aidge::xavierUniformFiller(std::shared_ptr<Aidge::Tensor> tensor,
+ T scaling, Aidge::VarianceNorm varianceNorm) {
AIDGE_ASSERT(tensor->getImpl(),
"Tensor got no implementation, cannot fill it.");
AIDGE_ASSERT(NativeType<T>::type == tensor->dataType(), "Wrong data type");
@@ -30,9 +31,6 @@ void Aidge::xavierUniformFiller(std::shared_ptr<Aidge::Tensor> tensor, T scaling
: fanOut);
const T scale(std::sqrt(3.0 / n));
- std::random_device rd;
- std::mt19937 gen(rd()); // Mersenne Twister pseudo-random number generator
-
std::uniform_real_distribution<T> uniformDist(-scale, scale);
std::shared_ptr<Aidge::Tensor> cpyTensor;
@@ -41,8 +39,8 @@ void Aidge::xavierUniformFiller(std::shared_ptr<Aidge::Tensor> tensor, T scaling
tensor->refCastFrom(cpyTensor, tensor->dataType(), "cpu");
// Setting values
for (std::size_t idx = 0; idx < tensorWithValues.size(); ++idx) {
- T value = scaling * uniformDist(gen);
- tensorWithValues.set<T>(idx, value);
+ tensorWithValues.set<T>(
+ idx, scaling * uniformDist(Aidge::Random::Generator::get()));
}
// Copy values back to the original tensors (actual copy only if needed)
@@ -50,7 +48,7 @@ void Aidge::xavierUniformFiller(std::shared_ptr<Aidge::Tensor> tensor, T scaling
}
template <typename T>
void Aidge::xavierNormalFiller(std::shared_ptr<Aidge::Tensor> tensor, T scaling,
- Aidge::VarianceNorm varianceNorm) {
+ Aidge::VarianceNorm varianceNorm) {
AIDGE_ASSERT(tensor->getImpl(),
"Tensor got no implementation, cannot fill it.");
AIDGE_ASSERT(NativeType<T>::type == tensor->dataType(), "Wrong data type");
@@ -64,9 +62,6 @@ void Aidge::xavierNormalFiller(std::shared_ptr<Aidge::Tensor> tensor, T scaling,
: fanOut);
const double stdDev(std::sqrt(1.0 / n));
- std::random_device rd;
- std::mt19937 gen(rd()); // Mersenne Twister pseudo-random number generator
-
std::normal_distribution<T> normalDist(0.0, stdDev);
std::shared_ptr<Aidge::Tensor> cpyTensor;
@@ -76,7 +71,8 @@ void Aidge::xavierNormalFiller(std::shared_ptr<Aidge::Tensor> tensor, T scaling,
// Setting values
for (std::size_t idx = 0; idx < tensorWithValues.size(); ++idx) {
- tensorWithValues.set<T>(idx, scaling*normalDist(gen));
+ tensorWithValues.set<T>(
+ idx, scaling * normalDist(Aidge::Random::Generator::get()));
}
// Copy values back to the original tensors (actual copy only if needed)
diff --git a/src/utils/Random.cpp b/src/utils/Random.cpp
new file mode 100644
index 000000000..e3716f4fc
--- /dev/null
+++ b/src/utils/Random.cpp
@@ -0,0 +1,22 @@
+/********************************************************************************
+ * Copyright (c) 2023 CEA-List
+ *
+ * This program and the accompanying materials are made available under the
+ * terms of the Eclipse Public License 2.0 which is available at
+ * http://www.eclipse.org/legal/epl-2.0.
+ *
+ * SPDX-License-Identifier: EPL-2.0
+ *
+ ********************************************************************************/
+
+#include "aidge/utils/Random.hpp"
+
+#include <random> // normal_distribution, uniform_real_distribution
+
+std::mt19937 Aidge::Random::Generator::generator{std::random_device{}()};
+unsigned int Aidge::Random::Generator::seed = 0;
+
+void Aidge::Random::Generator::setSeed(unsigned int seed) {
+ seed = seed;
+ generator.seed(seed);
+}
--
GitLab