chabisik/cppbasedann
1
0
Fork 0
Code Issues Pull requests Releases Wiki Activity

Adding ANN print function

Browse source
This commit is contained in:
chabisik 2022-01-09 14:56:35 +01:00
parent 26ae13b72e
commit 639694d05c
3 changed files with 44 additions and 46 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

6
main.cpp
View file

@ -4,7 +4,7 @@
#include "myclasses.h"
#include <vector>
#include <iterator>
using namespace std;
@ -26,6 +26,8 @@ int main(int argc, char *argv[])
n0.activate(it);
cout << "is = " << n0.get_output() << endl;*/
Network(4, 5);
Network network(4, 5);
network.print();
return 0;
}

75
myclasses.cpp
View file

@ -6,13 +6,12 @@
using namespace std;
Neuron::Neuron(int prev_layer_size, Activ activ_function)
Neuron::Neuron(int prev_layer_size)
{
for(int i(1) ; i<=prev_layer_size ; i++)
{
weights.push_front(Tools::get_random(0.0, 1.0));
}
activ = activ_function;
bias = 0.1;
output = 0.0;
derror = 0.0;
@ -23,16 +22,16 @@ void Neuron::set_output(float value)
output = value;
}
void Neuron::activate(forward_list<Neuron>::iterator &prev_layer_it)
void Neuron::activate(forward_list<Neuron>::iterator &prev_layer_it, Activ activ_function)
{
set_output(bias);
output = bias;
for(forward_list<float>::iterator it(weights.begin()) ; it!=weights.end() ; ++it)
{
output += (*it) * ((*prev_layer_it).output);
prev_layer_it++;
}
switch(activ)
switch(activ_function)
{
case RELU:
output = (output > 0.0) ? output : 0.0;
@ -66,13 +65,13 @@ Network::Network(int n_layers, int n_neurons)
{
if(i==1)
{
current_layer.push_front( Neuron(0, LINEAR) );
current_layer.push_front( Neuron(0) );
}else if(i==n_layers)
{
current_layer.push_front( Neuron(n_neurons, SIGMOID) );
current_layer.push_front( Neuron(n_neurons) );
}else
{
current_layer.push_front( Neuron(n_neurons, RELU) );
current_layer.push_front( Neuron(n_neurons) );
}
}
layers.push_back(current_layer);
@ -90,13 +89,13 @@ Network::Network(const std::vector<int> &n_neurons, Activ h_activ, Activ o_activ
{
if(i==0)
{
current_layer.push_front( Neuron(0, LINEAR) );
current_layer.push_front( Neuron(0) );
}else if(i==n_neurons.size()-1)
{
current_layer.push_front( Neuron(n_neurons[i-1], o_activ) );
current_layer.push_front( Neuron(n_neurons[i-1]) );
}else
{
current_layer.push_front( Neuron(n_neurons[i-1], h_activ) );
current_layer.push_front( Neuron(n_neurons[i-1]) );
}
}
layers.push_back(current_layer);
@ -112,50 +111,48 @@ void Network::print()
cout << "#>>==========================================<<#" << endl;
cout << ">> Number of layers : " << layers.size() << endl;
cout << "------------------------------------------------" << endl;
int layer_counter = 0;
int prev_layer_size_temp = 0, params_counter = 0;
for(list<forward_list<Neuron>>::iterator it1(layers.begin()) ; it1!=layers.end() ; ++it1)
{
layer_counter++;
int current_layer_size = 0;
for(forward_list<Neuron>::iterator it2(it1) ; it2!=it1.end() ; ++it2)
for(forward_list<Neuron>::iterator it2(it1->begin()) ; it2!=it1->end() ; ++it2)
{
current_layer_size++;
}
if(i==0)
if(layer_counter==1)
{
prev_layer_size_temp = current_layer_size;
}
else
{
params_counter += (prev_layer_size_temp+1)*current_layer_size;
prev_layer_size_temp = current_layer_size;
}
if(layer_counter==1)
{
cout << ">> Input layer" << endl;
cout << "size : " << layers << endl;
cout << "size : " << current_layer_size << endl;
cout << "neurons' outputs : ";
temp = network->layers_first_neurons[i];
while(temp != NULL)
{
cout << ("%f ", temp->output);
temp = temp->same_layer_next_neuron;
}
cout << ("\n");
}else if(i==layers.size()-1)
for(forward_list<Neuron>::iterator it2(it1->begin()) ; it2!=it1->end() ; ++it2){it2->get_output();}
cout << endl;
}else if(layer_counter==layers.size())
{
cout << (">> Output layer\n");
cout << ("size : %d\n", network->neurons_per_layer[i]);
cout << "size : " << current_layer_size << endl;
cout << ("neurons' outputs : ");
temp = network->layers_first_neurons[i];
while(temp != NULL)
{
cout << ("%f ", temp->output);
temp = temp->same_layer_next_neuron;
}
cout << ("\n");
for(forward_list<Neuron>::iterator it2(it1->begin()) ; it2!=it1->end() ; ++it2){it2->get_output();}
cout << endl;
}else
{
cout << (">> Hidden layer %d\n", i);
cout << ("size : %d\n", network->neurons_per_layer[i]);
cout << ">> Hidden layer " << layer_counter-1 << endl;
cout << "size : " << current_layer_size << endl;
}
cout << ("------------------------------------------------\n");
cout << "------------------------------------------------" << endl;
}
cout << ("Number of parameters : ");
for(i=1 ; i<network->n_layers ; i++)
{
n_params += network->neurons_per_layer[i] * (network->neurons_per_layer[i-1] + 1);
}
cout << ("%d\n", n_params);
cout << "Number of parameters : ";
cout << params_counter << endl;
cout << "#>>==========================================<<#" << endl << endl;
}

9
myclasses.h
View file

@ -8,22 +8,21 @@
enum Activ
{
RELU, TANH, SIGMOID, LINEAR
RELU, TANH, SIGMOID, LINEAR, SOFTMAX
};
class Neuron
{
public:
Neuron(int prev_layer_size, Activ activ_function);
Neuron(int prev_layer_size); //prev_layer_size = number of weights
void set_output(float value);
float get_output();//to be deleted
void activate(std::forward_list<Neuron>::iterator &prev_layer_it);
void activate(std::forward_list<Neuron>::iterator &prev_layer_it, Activ activ_function=LINEAR);
private:
std::forward_list<float> weights;
float bias;
float output;
float derror;
Activ activ;
};
@ -32,7 +31,7 @@ class Network
public:
Network(int n_layers, int n_neurons);
Network(const std::vector<int> &n_neurons, Activ h_activ=RELU, Activ o_activ=SIGMOID);
void print() const;
void print();
bool forward(const std::vector<float> &input, const std::vector<float> &target);
bool backward();
private: