mirror of
https://github.com/Relintai/pmlpp.git
synced 2024-12-21 14:56:47 +01:00
Now MLPPTanhReg uses engine classes.
This commit is contained in:
parent
d6037730b5
commit
aa8043621e
@ -12,7 +12,6 @@
|
||||
#include "../regularization/reg.h"
|
||||
#include "../utilities/utilities.h"
|
||||
|
||||
#include <iostream>
|
||||
#include <random>
|
||||
|
||||
/*
|
||||
@ -62,11 +61,14 @@ void MLPPTanhReg::set_alpha(const real_t val) {
|
||||
}
|
||||
*/
|
||||
|
||||
std::vector<real_t> MLPPTanhReg::model_set_test(std::vector<std::vector<real_t>> X) {
|
||||
// Ref<MLPPVector> model_set_test(const Ref<MLPPMatrix> &X);
|
||||
// real_t model_test(const Ref<MLPPVector> &x);
|
||||
|
||||
Ref<MLPPVector> MLPPTanhReg::model_set_test(const Ref<MLPPMatrix> &X) {
|
||||
return evaluatem(X);
|
||||
}
|
||||
|
||||
real_t MLPPTanhReg::model_test(std::vector<real_t> x) {
|
||||
real_t MLPPTanhReg::model_test(const Ref<MLPPVector> &x) {
|
||||
return evaluatev(x);
|
||||
}
|
||||
|
||||
@ -83,21 +85,21 @@ void MLPPTanhReg::gradient_descent(real_t learning_rate, int max_epoch, bool ui)
|
||||
while (true) {
|
||||
cost_prev = cost(_y_hat, _output_set);
|
||||
|
||||
std::vector<real_t> error = alg.subtraction(_y_hat, _output_set);
|
||||
Ref<MLPPVector> error = alg.subtractionnv(_y_hat, _output_set);
|
||||
|
||||
_weights = alg.subtraction(_weights, alg.scalarMultiply(learning_rate / _n, alg.mat_vec_mult(alg.transpose(_input_set), alg.hadamard_product(error, avn.tanh(_z, 1)))));
|
||||
_weights = alg.subtractionnv(_weights, alg.scalar_multiplynv(learning_rate / _n, alg.mat_vec_multv(alg.transposem(_input_set), alg.hadamard_productnv(error, avn.tanh_derivv(_z)))));
|
||||
//_reg
|
||||
_weights = regularization.regWeights(_weights, _lambda, _alpha, "None");
|
||||
_weights = regularization.reg_weightsv(_weights, _lambda, _alpha, MLPPReg::REGULARIZATION_TYPE_NONE);
|
||||
|
||||
// Calculating the bias gradients
|
||||
_bias -= learning_rate * alg.sum_elements(alg.hadamard_product(error, avn.tanh(_z, 1))) / _n;
|
||||
_bias -= learning_rate * alg.sum_elementsv(alg.hadamard_productnv(error, avn.tanh_derivv(_z))) / _n;
|
||||
|
||||
forward_pass();
|
||||
|
||||
// UI PORTION
|
||||
if (ui) {
|
||||
MLPPUtilities::CostInfo(epoch, cost_prev, cost(_y_hat, _output_set));
|
||||
MLPPUtilities::UI(_weights, _bias);
|
||||
MLPPUtilities::cost_info(epoch, cost_prev, cost(_y_hat, _output_set));
|
||||
MLPPUtilities::print_ui_vb(_weights, _bias);
|
||||
}
|
||||
|
||||
epoch++;
|
||||
@ -119,28 +121,47 @@ void MLPPTanhReg::sgd(real_t learning_rate, int max_epoch, bool ui) {
|
||||
std::default_random_engine generator(rd());
|
||||
std::uniform_int_distribution<int> distribution(0, int(_n - 1));
|
||||
|
||||
Ref<MLPPVector> input_set_row_tmp;
|
||||
input_set_row_tmp.instance();
|
||||
input_set_row_tmp->resize(_input_set->size().x);
|
||||
|
||||
Ref<MLPPVector> output_set_row_tmp;
|
||||
output_set_row_tmp.instance();
|
||||
output_set_row_tmp->resize(1);
|
||||
|
||||
Ref<MLPPVector> y_hat_row_tmp;
|
||||
y_hat_row_tmp.instance();
|
||||
y_hat_row_tmp->resize(1);
|
||||
|
||||
while (true) {
|
||||
int outputIndex = distribution(generator);
|
||||
int output_index = distribution(generator);
|
||||
|
||||
real_t y_hat = evaluatev(_input_set[outputIndex]);
|
||||
cost_prev = cost({ _y_hat }, { _output_set[outputIndex] });
|
||||
_input_set->get_row_into_mlpp_vector(output_index, input_set_row_tmp);
|
||||
real_t output_set_entry = _output_set->get_element(output_index);
|
||||
output_set_row_tmp->set_element(0, output_set_entry);
|
||||
|
||||
real_t error = y_hat - _output_set[outputIndex];
|
||||
real_t y_hat = evaluatev(input_set_row_tmp);
|
||||
y_hat_row_tmp->set_element(0, y_hat);
|
||||
|
||||
cost_prev = cost(y_hat_row_tmp, output_set_row_tmp);
|
||||
|
||||
real_t error = y_hat - output_set_entry;
|
||||
|
||||
// Weight Updation
|
||||
_weights = alg.subtraction(_weights, alg.scalarMultiply(learning_rate * error * (1 - y_hat * y_hat), _input_set[outputIndex]));
|
||||
_weights = alg.subtractionnv(_weights, alg.scalar_multiplynv(learning_rate * error * (1 - y_hat * y_hat), input_set_row_tmp));
|
||||
//_reg
|
||||
_weights = regularization.regWeights(_weights, _lambda, _alpha, "None");
|
||||
_weights = regularization.reg_weightsv(_weights, _lambda, _alpha, MLPPReg::REGULARIZATION_TYPE_NONE);
|
||||
|
||||
// Bias updation
|
||||
_bias -= learning_rate * error * (1 - y_hat * y_hat);
|
||||
|
||||
y_hat = evaluatev(_input_set[outputIndex]);
|
||||
y_hat = evaluatev(input_set_row_tmp);
|
||||
|
||||
if (ui) {
|
||||
MLPPUtilities::CostInfo(epoch, cost_prev, cost({ _y_hat }, { _output_set[outputIndex] }));
|
||||
MLPPUtilities::UI(_weights, _bias);
|
||||
MLPPUtilities::cost_info(epoch, cost_prev, cost(y_hat_row_tmp, output_set_row_tmp));
|
||||
MLPPUtilities::print_ui_vb(_weights, _bias);
|
||||
}
|
||||
|
||||
epoch++;
|
||||
|
||||
if (epoch > max_epoch) {
|
||||
@ -161,33 +182,34 @@ void MLPPTanhReg::mbgd(real_t learning_rate, int max_epoch, int mini_batch_size,
|
||||
|
||||
// Creating the mini-batches
|
||||
int n_mini_batch = _n / mini_batch_size;
|
||||
auto batches = MLPPUtilities::createMiniBatches(_input_set, _output_set, n_mini_batch);
|
||||
auto inputMiniBatches = std::get<0>(batches);
|
||||
auto outputMiniBatches = std::get<1>(batches);
|
||||
MLPPUtilities::CreateMiniBatchMVBatch batches = MLPPUtilities::create_mini_batchesmv(_input_set, _output_set, n_mini_batch);
|
||||
|
||||
while (true) {
|
||||
for (int i = 0; i < n_mini_batch; i++) {
|
||||
std::vector<real_t> y_hat = evaluatem(inputMiniBatches[i]);
|
||||
std::vector<real_t> z = propagatem(inputMiniBatches[i]);
|
||||
cost_prev = cost(y_hat, outputMiniBatches[i]);
|
||||
Ref<MLPPMatrix> current_input_batch_entry = batches.input_sets[i];
|
||||
Ref<MLPPVector> current_output_batch_entry = batches.output_sets[i];
|
||||
|
||||
std::vector<real_t> error = alg.subtraction(y_hat, outputMiniBatches[i]);
|
||||
Ref<MLPPVector> y_hat = evaluatem(current_input_batch_entry);
|
||||
Ref<MLPPVector> z = propagatem(current_input_batch_entry);
|
||||
cost_prev = cost(y_hat, current_output_batch_entry);
|
||||
|
||||
Ref<MLPPVector> error = alg.subtractionnv(y_hat, current_output_batch_entry);
|
||||
|
||||
// Calculating the weight gradients
|
||||
_weights = alg.subtraction(_weights, alg.scalarMultiply(learning_rate / _n, alg.mat_vec_mult(alg.transpose(inputMiniBatches[i]), alg.hadamard_product(error, avn.tanh(z, 1)))));
|
||||
_weights = alg.subtractionnv(_weights, alg.scalar_multiplynv(learning_rate / _n, alg.mat_vec_multv(alg.transposem(current_input_batch_entry), alg.hadamard_productnv(error, avn.tanh_derivv(z)))));
|
||||
//_reg
|
||||
_weights = regularization.regWeights(_weights, _lambda, _alpha, "None");
|
||||
_weights = regularization.reg_weightsv(_weights, _lambda, _alpha, MLPPReg::REGULARIZATION_TYPE_NONE);
|
||||
|
||||
// Calculating the bias gradients
|
||||
_bias -= learning_rate * alg.sum_elements(alg.hadamard_product(error, avn.tanh(_z, true))) / _n;
|
||||
_bias -= learning_rate * alg.sum_elementsv(alg.hadamard_productnv(error, avn.tanh_derivv(_z))) / _n;
|
||||
|
||||
forward_pass();
|
||||
|
||||
y_hat = evaluatem(inputMiniBatches[i]);
|
||||
y_hat = evaluatem(current_input_batch_entry);
|
||||
|
||||
if (ui) {
|
||||
MLPPUtilities::CostInfo(epoch, cost_prev, cost(y_hat, outputMiniBatches[i]));
|
||||
MLPPUtilities::UI(_weights, _bias);
|
||||
MLPPUtilities::cost_info(epoch, cost_prev, cost(y_hat, current_output_batch_entry));
|
||||
MLPPUtilities::print_ui_vb(_weights, _bias);
|
||||
}
|
||||
}
|
||||
|
||||
@ -204,13 +226,13 @@ void MLPPTanhReg::mbgd(real_t learning_rate, int max_epoch, int mini_batch_size,
|
||||
real_t MLPPTanhReg::score() {
|
||||
MLPPUtilities util;
|
||||
|
||||
return util.performance(_y_hat, _output_set);
|
||||
return util.performance_vec(_y_hat, _output_set);
|
||||
}
|
||||
|
||||
void MLPPTanhReg::save(std::string file_name) {
|
||||
MLPPUtilities util;
|
||||
void MLPPTanhReg::save(const String &file_name) {
|
||||
//MLPPUtilities util;
|
||||
|
||||
util.saveParameters(file_name, _weights, _bias);
|
||||
//util.saveParameters(file_name, _weights, _bias);
|
||||
}
|
||||
|
||||
bool MLPPTanhReg::is_initialized() {
|
||||
@ -226,53 +248,68 @@ void MLPPTanhReg::initialize() {
|
||||
_initialized = true;
|
||||
}
|
||||
|
||||
MLPPTanhReg::MLPPTanhReg(std::vector<std::vector<real_t>> p_input_set, std::vector<real_t> p_output_set, MLPPReg::RegularizationType p_reg, real_t p_lambda, real_t p_alpha) {
|
||||
MLPPTanhReg::MLPPTanhReg(const Ref<MLPPMatrix> &p_input_set, const Ref<MLPPVector> &p_output_set, MLPPReg::RegularizationType p_reg, real_t p_lambda, real_t p_alpha) {
|
||||
_input_set = p_input_set;
|
||||
_output_set = p_output_set;
|
||||
_n = _input_set.size();
|
||||
_k = _input_set[0].size();
|
||||
_n = _input_set->size().y;
|
||||
_k = _input_set->size().x;
|
||||
_reg = p_reg;
|
||||
_lambda = p_lambda;
|
||||
_alpha = p_alpha;
|
||||
|
||||
_y_hat.resize(_n);
|
||||
_weights = MLPPUtilities::weightInitialization(_k);
|
||||
_bias = MLPPUtilities::biasInitialization();
|
||||
_y_hat.instance();
|
||||
_y_hat->resize(_n);
|
||||
|
||||
MLPPUtilities utils;
|
||||
|
||||
_weights.instance();
|
||||
_weights->resize(_k);
|
||||
|
||||
utils.weight_initializationv(_weights);
|
||||
|
||||
_bias = utils.bias_initializationr();
|
||||
|
||||
_initialized = true;
|
||||
}
|
||||
|
||||
MLPPTanhReg::MLPPTanhReg() {
|
||||
_initialized = false;
|
||||
}
|
||||
MLPPTanhReg::~MLPPTanhReg() {
|
||||
}
|
||||
|
||||
real_t MLPPTanhReg::cost(std::vector<real_t> y_hat, std::vector<real_t> y) {
|
||||
real_t MLPPTanhReg::cost(const Ref<MLPPVector> &y_hat, const Ref<MLPPVector> &y) {
|
||||
MLPPReg regularization;
|
||||
class MLPPCost cost;
|
||||
MLPPCost mlpp_cost;
|
||||
|
||||
//_reg
|
||||
return cost.MSE(y_hat, y) + regularization.regTerm(_weights, _lambda, _alpha, "None");
|
||||
return mlpp_cost.msev(y_hat, y) + regularization.reg_termv(_weights, _lambda, _alpha, MLPPReg::REGULARIZATION_TYPE_NONE);
|
||||
}
|
||||
|
||||
real_t MLPPTanhReg::evaluatev(std::vector<real_t> x) {
|
||||
real_t MLPPTanhReg::evaluatev(const Ref<MLPPVector> &x) {
|
||||
MLPPLinAlg alg;
|
||||
MLPPActivation avn;
|
||||
return avn.tanh(alg.dot(_weights, x) + _bias);
|
||||
|
||||
return avn.tanh_normr(alg.dotv(_weights, x) + _bias);
|
||||
}
|
||||
|
||||
real_t MLPPTanhReg::propagatev(std::vector<real_t> x) {
|
||||
real_t MLPPTanhReg::propagatev(const Ref<MLPPVector> &x) {
|
||||
MLPPLinAlg alg;
|
||||
return alg.dot(_weights, x) + _bias;
|
||||
|
||||
return alg.dotv(_weights, x) + _bias;
|
||||
}
|
||||
|
||||
std::vector<real_t> MLPPTanhReg::evaluatem(std::vector<std::vector<real_t>> X) {
|
||||
Ref<MLPPVector> MLPPTanhReg::evaluatem(const Ref<MLPPMatrix> &X) {
|
||||
MLPPLinAlg alg;
|
||||
MLPPActivation avn;
|
||||
return avn.tanh(alg.scalarAdd(_bias, alg.mat_vec_mult(X, _weights)));
|
||||
|
||||
return avn.tanh_normv(alg.scalar_addnv(_bias, alg.mat_vec_multv(X, _weights)));
|
||||
}
|
||||
|
||||
std::vector<real_t> MLPPTanhReg::propagatem(std::vector<std::vector<real_t>> X) {
|
||||
Ref<MLPPVector> MLPPTanhReg::propagatem(const Ref<MLPPMatrix> &X) {
|
||||
MLPPLinAlg alg;
|
||||
return alg.scalarAdd(_bias, alg.mat_vec_mult(X, _weights));
|
||||
|
||||
return alg.scalar_addnv(_bias, alg.mat_vec_multv(X, _weights));
|
||||
}
|
||||
|
||||
// Tanh ( wTx + b )
|
||||
@ -280,7 +317,7 @@ void MLPPTanhReg::forward_pass() {
|
||||
MLPPActivation avn;
|
||||
|
||||
_z = propagatem(_input_set);
|
||||
_y_hat = avn.tanh(_z);
|
||||
_y_hat = avn.tanh_normv(_z);
|
||||
}
|
||||
|
||||
void MLPPTanhReg::_bind_methods() {
|
||||
|
||||
@ -17,9 +17,6 @@
|
||||
|
||||
#include "../regularization/reg.h"
|
||||
|
||||
#include <string>
|
||||
#include <vector>
|
||||
|
||||
class MLPPTanhReg : public Reference {
|
||||
GDCLASS(MLPPTanhReg, Reference);
|
||||
|
||||
@ -41,8 +38,8 @@ public:
|
||||
void set_alpha(const real_t val);
|
||||
*/
|
||||
|
||||
std::vector<real_t> model_set_test(std::vector<std::vector<real_t>> X);
|
||||
real_t model_test(std::vector<real_t> x);
|
||||
Ref<MLPPVector> model_set_test(const Ref<MLPPMatrix> &X);
|
||||
real_t model_test(const Ref<MLPPVector> &x);
|
||||
|
||||
void gradient_descent(real_t learning_rate, int max_epoch, bool ui = false);
|
||||
void sgd(real_t learning_rate, int max_epoch, bool ui = false);
|
||||
@ -50,34 +47,34 @@ public:
|
||||
|
||||
real_t score();
|
||||
|
||||
void save(std::string file_name);
|
||||
void save(const String &file_name);
|
||||
|
||||
bool is_initialized();
|
||||
void initialize();
|
||||
|
||||
MLPPTanhReg(std::vector<std::vector<real_t>> p_input_set, std::vector<real_t> p_output_set, MLPPReg::RegularizationType p_reg = MLPPReg::REGULARIZATION_TYPE_NONE, real_t p_lambda = 0.5, real_t p_alpha = 0.5);
|
||||
MLPPTanhReg(const Ref<MLPPMatrix> &p_input_set, const Ref<MLPPVector> &p_output_set, MLPPReg::RegularizationType p_reg = MLPPReg::REGULARIZATION_TYPE_NONE, real_t p_lambda = 0.5, real_t p_alpha = 0.5);
|
||||
|
||||
MLPPTanhReg();
|
||||
~MLPPTanhReg();
|
||||
|
||||
protected:
|
||||
real_t cost(std::vector<real_t> y_hat, std::vector<real_t> y);
|
||||
real_t cost(const Ref<MLPPVector> &y_hat, const Ref<MLPPVector> &y);
|
||||
|
||||
real_t evaluatev(std::vector<real_t> x);
|
||||
real_t propagatev(std::vector<real_t> x);
|
||||
real_t evaluatev(const Ref<MLPPVector> &x);
|
||||
real_t propagatev(const Ref<MLPPVector> &x);
|
||||
|
||||
std::vector<real_t> evaluatem(std::vector<std::vector<real_t>> X);
|
||||
std::vector<real_t> propagatem(std::vector<std::vector<real_t>> X);
|
||||
Ref<MLPPVector> evaluatem(const Ref<MLPPMatrix> &X);
|
||||
Ref<MLPPVector> propagatem(const Ref<MLPPMatrix> &X);
|
||||
|
||||
void forward_pass();
|
||||
|
||||
static void _bind_methods();
|
||||
|
||||
std::vector<std::vector<real_t>> _input_set;
|
||||
std::vector<real_t> _output_set;
|
||||
std::vector<real_t> _z;
|
||||
std::vector<real_t> _y_hat;
|
||||
std::vector<real_t> _weights;
|
||||
Ref<MLPPMatrix> _input_set;
|
||||
Ref<MLPPVector> _output_set;
|
||||
Ref<MLPPVector> _z;
|
||||
Ref<MLPPVector> _y_hat;
|
||||
Ref<MLPPVector> _weights;
|
||||
real_t _bias;
|
||||
|
||||
int _n;
|
||||
|
||||
Loading…
Reference in New Issue
Block a user