Skip to content
Snippets Groups Projects
Commit e362ef7a authored by Martin Beseda's avatar Martin Beseda
Browse files

FIX: Merge

parent f2a604d9
No related branches found
No related tags found
No related merge requests found
Hide whitespace changes
Inline Side-by-side
src/DataSet/DataSet.cpp
+ 7
− 11
View file @ e362ef7a
......@@ -17,13 +17,10 @@ namespace lib4neuro {
this->n_elements = 0;
this->input_dim = 0;
this->output_dim = 0;
this->gen = boost::random::mt19937(std::time(0));
}
DataSet::DataSet(std::string file_path) {
std::ifstream ifs(file_path);
this->gen = boost::random::mt19937(std::time(0));
if(ifs.is_open()) {
try {
boost::archive::text_iarchive ia(ifs);
......@@ -45,7 +42,6 @@ namespace lib4neuro {
this->data = *data_ptr;
this->input_dim = this->data[0].first.size();
this->output_dim = this->data[0].second.size();
this->gen = boost::random::mt19937(std::time(0));
if(ns) {
this->normalization_strategy = ns;
......@@ -66,7 +62,6 @@ namespace lib4neuro {
this->n_elements = 0;
this->input_dim = 1;
this->output_dim = 1;
this->gen = boost::random::mt19937(std::time(0));
if(ns) {
this->normalization_strategy = ns;
......@@ -87,7 +82,6 @@ namespace lib4neuro {
this->input_dim = bounds.size() / 2;
this->output_dim = output_dim;
this->n_elements = 0;
this->gen = boost::random::mt19937(std::time(0));
if(ns) {
this->normalization_strategy = ns;
......@@ -152,7 +146,7 @@ namespace lib4neuro {
}
for (unsigned int i = 0; i < bounds.size(); i += 2) {
if (no_elems_in_one_dim == 1) {
if (no_elems_in_one_dim == 1) {
frac = 1;
} else {
frac = (bounds[i] - bounds[i+1]) / (no_elems_in_one_dim - 1);
......@@ -406,17 +400,19 @@ namespace lib4neuro {
* Method returning random amount of data pairs between 1-max
*/
std::vector<std::pair<std::vector<double>, std::vector<double>>> DataSet::get_random_data_batch(size_t max) {
if (max <= 0 || max >= this->data.size()) {
if (max <= 0) {
return this->data;
} else {
std::vector<std::pair<std::vector<double>, std::vector<double>>> newData;
boost::random::uniform_int_distribution<> dist(0, this->data.size() - 1);
srand(time(NULL)); //TODO use Mersen twister from Boost
size_t n_chosen = rand() % std::min(max, this->data.size())+1;
n_chosen = max;
std::vector<size_t> chosens;
size_t chosen;
for (int i = 0; i < max; i++) {
chosen = dist(gen);
for (int i = 0; i < n_chosen; i++) {
chosen = rand() % this->data.size();
auto it = std::find(chosens.begin(), chosens.end(), chosen);
if (it != chosens.end()) {
......
src/DataSet/DataSet.h
+ 1
− 6
View file @ e362ef7a
......@@ -12,9 +12,6 @@
#include <string>
#include <functional>
#include <limits>
#include <boost/random/mersenne_twister.hpp>
#include <boost/random/uniform_int_distribution.hpp>
#include <ctime>
#include "../settings.h"
#include "../NormalizationStrategy/NormalizationStrategy.h"
......@@ -29,8 +26,6 @@ namespace lib4neuro {
private:
boost::random::mt19937 gen;
/**
* Number of elements in the data set
*/
......@@ -293,7 +288,7 @@ namespace lib4neuro {
* @param max
* @return
*/
LIB4NEURO_API std::vector<std::pair<std::vector<double>, std::vector<double>>> get_random_data_batch(size_t max);
LIB4NEURO_API std::vector<std::pair<std::vector<double>, std::vector<double>>> get_random_data_batch(size_t max);
};
}
#endif //INC_4NEURO_DATASET_H
src/Network/NeuralNetwork.cpp
+ 29
− 32
View file @ e362ef7a
......@@ -35,8 +35,6 @@ namespace lib4neuro {
this->delete_weights = true;
this->delete_biases = true;
this->layers_analyzed = false;
this->gen = boost::random::mt19937(std::time(0));
}
NeuralNetwork::NeuralNetwork(std::string filepath) {
......@@ -547,30 +545,6 @@ namespace lib4neuro {
this->delete_weights = false;
}
void NeuralNetwork::get_jacobian(std::vector<std::vector<double>> &jacobian, std::pair<std::vector<double>, std::vector<double>> &data, std::vector<double> &error) {
std::vector<double> fv(this->get_n_outputs());
jacobian.resize(this->get_n_outputs());
error.resize(this->get_n_outputs());
for(size_t i = 0; i < this->get_n_outputs(); ++i){
jacobian[i].resize(this->get_n_weights() + this->get_n_biases());
std::fill(jacobian[i].begin(), jacobian[i].end(), 0);
}
this->eval_single( data.first, fv );
std::vector<double> error_partial(this->get_n_outputs());
std::fill(error_partial.begin(), error_partial.end(), 0.0);
for( size_t i = 0; i < this->get_n_outputs(); ++i){
error_partial[i] = 1;
this->add_to_gradient_single(data.first, error_partial, 1.0, jacobian[i]);
error[i] = data.second[i] - fv[i];
error_partial[i] = 0;
}
}
void NeuralNetwork::eval_single(::std::vector<double>& input,
::std::vector<double>& output,
::std::vector<double>* custom_weights_and_biases) {
......@@ -789,7 +763,7 @@ namespace lib4neuro {
void NeuralNetwork::randomize_weights() {
boost::random::mt19937 gen(std::time(0));
// Init weight guess ("optimal" for logistic activation functions)
double r = 4 * sqrt(6. / (this->connection_weights->size()));
......@@ -803,7 +777,7 @@ namespace lib4neuro {
void NeuralNetwork::randomize_biases() {
boost::random::mt19937 gen(std::time(0));
// Init weight guess ("optimal" for logistic activation functions)
boost::random::uniform_real_distribution<> dist(-1, 1);
......@@ -1226,8 +1200,6 @@ namespace lib4neuro {
"doesn't specify input and output layers, which are compulsory!");
}
this->gen = boost::random::mt19937(std::time(0));
this->neurons = new ::std::vector<Neuron *>(0);
this->neuron_biases = new ::std::vector<double>(0);
this->neuron_potentials = new ::std::vector<double>(0);
......@@ -1256,8 +1228,8 @@ namespace lib4neuro {
COUT_DEBUG("# of outputs: " << out_dim << std::endl);
WRITE_TO_OFS_DEBUG(ofs, "Fully connected feed-forward network being constructed:" << std::endl
<< "# of inputs: " << inp_dim << std::endl
<< "# of outputs: " << out_dim << std::endl);
<< "# of inputs: " << inp_dim << std::endl
<< "# of outputs: " << out_dim << std::endl);
std::vector<size_t> input_layer_neuron_indices;
std::vector<size_t> previous_layer_neuron_indices;
......@@ -1348,5 +1320,30 @@ namespace lib4neuro {
this->analyze_layer_structure();
}
void NeuralNetwork::get_jacobian(std::vector<std::vector<double>> &jacobian, std::pair<std::vector<double>, std::vector<double>> &data, std::vector<double> &error) {
std::vector<double> fv(this->get_n_outputs());
jacobian.resize(this->get_n_outputs());
error.resize(this->get_n_outputs());
for(size_t i = 0; i < this->get_n_outputs(); ++i){
jacobian[i].resize(this->get_n_weights() + this->get_n_biases());
std::fill(jacobian[i].begin(), jacobian[i].end(), 0);
}
this->eval_single( data.first, fv );
std::vector<double> error_partial(this->get_n_outputs());
std::fill(error_partial.begin(), error_partial.end(), 0.0);
for( size_t i = 0; i < this->get_n_outputs(); ++i){
error_partial[i] = 1;
this->add_to_gradient_single(data.first, error_partial, 1.0, jacobian[i]);
error[i] = data.second[i] - fv[i];
error_partial[i] = 0;
}
}
}
src/Network/NeuralNetwork.h
+ 13
− 18
View file @ e362ef7a
......@@ -17,10 +17,6 @@
#include <algorithm>
#include <utility>
#include <fstream>
#include <boost/random/mersenne_twister.hpp>
#include <boost/random/uniform_real_distribution.hpp>
#include <ctime>
#include "../settings.h"
#include "../Neuron/Neuron.h"
......@@ -32,8 +28,6 @@
#include "../NetConnection/ConnectionFunctionIdentity.h"
#include "../NormalizationStrategy/NormalizationStrategy.h"
namespace lib4neuro {
/**
......@@ -55,9 +49,8 @@ namespace lib4neuro {
*
*/
class NeuralNetwork {
protected:
boost::random::mt19937 gen;
/**
*
*/
......@@ -169,6 +162,18 @@ namespace lib4neuro {
public:
/**
* Runs @data through the network and then computes partial derivatives with respect to each output function and adds them
* to seperate vectors in @jacobian. Also computes the out error and stores in the vector @error
* @param[out] jacobian
* @param[in] data
* @param[out] error
*/
LIB4NEURO_API virtual void
get_jacobian(std::vector<std::vector<double>> &jacobian, std::pair<std::vector<double>, std::vector<double>> &data, std::vector<double> &error);
/**
*
* @param input
......@@ -279,16 +284,6 @@ namespace lib4neuro {
LIB4NEURO_API void
add_existing_connection(size_t n1_idx, size_t n2_idx, size_t connection_idx, NeuralNetwork &parent_network);
/**
* Runs @data through the network and then computes partial derivatives with respect to each output function and adds them
* to seperate vectors in @jacobian. Also computes the out error and stores in the vector @error
* @param[out] jacobian
* @param[in] data
* @param[out] error
*/
LIB4NEURO_API virtual void
get_jacobian(std::vector<std::vector<double>> &jacobian, std::pair<std::vector<double>, std::vector<double>> &data, std::vector<double> &error);
/**
*
*/
......
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment