Relintai/pmlpp
1
0
Fork 0
mirror of https://github.com/Relintai/pmlpp.git synced 2024-12-22 15:06:47 +01:00
Code Issues Packages Projects Releases Wiki Activity

MLPPSVC cleanup.

Browse Source
This commit is contained in:
Relintai 2023-04-28 20:11:05 +02:00
parent 63d8dbf676
commit 19c9107309
1 changed files with 11 additions and 17 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

28
mlpp/svc/svc.cpp
View File

@ -58,7 +58,6 @@ void MLPPSVC::gradient_descent(real_t learning_rate, int max_epoch, bool ui) {
MLPPCost mlpp_cost; MLPPCost mlpp_cost;
MLPPActivation avn; MLPPActivation avn;
MLPPLinAlg alg;
MLPPReg regularization; MLPPReg regularization;
real_t cost_prev = 0; real_t cost_prev = 0;
@ -69,11 +68,11 @@ void MLPPSVC::gradient_descent(real_t learning_rate, int max_epoch, bool ui) {
while (true) { while (true) {
cost_prev = cost(_y_hat, _output_set, _weights, _c); cost_prev = cost(_y_hat, _output_set, _weights, _c);
_weights = alg.subtractionnv(_weights, alg.scalar_multiplynv(learning_rate / _n, alg.mat_vec_multnv(alg.transposenm(_input_set), mlpp_cost.hinge_loss_derivwv(_z, _output_set, _c)))); _weights->sub(_input_set->transposen()->mult_vec(mlpp_cost.hinge_loss_derivwv(_z, _output_set, _c))->scalar_multiplyn(learning_rate / _n));
_weights = regularization.reg_weightsv(_weights, learning_rate / _n, 0, MLPPReg::REGULARIZATION_TYPE_RIDGE); _weights = regularization.reg_weightsv(_weights, learning_rate / _n, 0, MLPPReg::REGULARIZATION_TYPE_RIDGE);
// Calculating the bias gradients // Calculating the bias gradients
_bias += learning_rate * alg.sum_elementsv(mlpp_cost.hinge_loss_derivwv(_y_hat, _output_set, _c)) / _n; _bias += learning_rate * mlpp_cost.hinge_loss_derivwv(_y_hat, _output_set, _c)->sum_elements() / _n;
forward_pass(); forward_pass();
@ -96,7 +95,6 @@ void MLPPSVC::sgd(real_t learning_rate, int max_epoch, bool ui) {
MLPPCost mlpp_cost; MLPPCost mlpp_cost;
MLPPActivation avn; MLPPActivation avn;
MLPPLinAlg alg;
MLPPReg regularization; MLPPReg regularization;
std::random_device rd; std::random_device rd;
@ -140,7 +138,7 @@ void MLPPSVC::sgd(real_t learning_rate, int max_epoch, bool ui) {
real_t cost_deriv = cost_deriv_vec->get_element(0); real_t cost_deriv = cost_deriv_vec->get_element(0);
// Weight Updation // Weight Updation
_weights = alg.subtractionnv(_weights, alg.scalar_multiplynv(learning_rate * cost_deriv, input_set_row_tmp)); _weights->sub(input_set_row_tmp->scalar_multiplyn(learning_rate * cost_deriv));
_weights = regularization.reg_weightsv(_weights, learning_rate, 0, MLPPReg::REGULARIZATION_TYPE_RIDGE); _weights = regularization.reg_weightsv(_weights, learning_rate, 0, MLPPReg::REGULARIZATION_TYPE_RIDGE);
// Bias updation // Bias updation
@ -168,7 +166,6 @@ void MLPPSVC::mbgd(real_t learning_rate, int max_epoch, int mini_batch_size, boo
MLPPCost mlpp_cost; MLPPCost mlpp_cost;
MLPPActivation avn; MLPPActivation avn;
MLPPLinAlg alg;
MLPPReg regularization; MLPPReg regularization;
real_t cost_prev = 0; real_t cost_prev = 0;
@ -190,11 +187,11 @@ void MLPPSVC::mbgd(real_t learning_rate, int max_epoch, int mini_batch_size, boo
cost_prev = cost(z, current_output_batch_entry, _weights, _c); cost_prev = cost(z, current_output_batch_entry, _weights, _c);
// Calculating the weight gradients // Calculating the weight gradients
_weights = alg.subtractionnv(_weights, alg.scalar_multiplynv(learning_rate / _n, alg.mat_vec_multnv(alg.transposenm(current_input_batch_entry), mlpp_cost.hinge_loss_derivwv(z, current_output_batch_entry, _c)))); _weights->subn(current_input_batch_entry->transposen()->mult_vec(mlpp_cost.hinge_loss_derivwv(z, current_output_batch_entry, _c))->scalar_multiplyn(learning_rate / _n));
_weights = regularization.reg_weightsv(_weights, learning_rate / _n, 0, MLPPReg::REGULARIZATION_TYPE_RIDGE); _weights = regularization.reg_weightsv(_weights, learning_rate / _n, 0, MLPPReg::REGULARIZATION_TYPE_RIDGE);
// Calculating the bias gradients // Calculating the bias gradients
_bias -= learning_rate * alg.sum_elementsv(mlpp_cost.hinge_loss_derivwv(y_hat, current_output_batch_entry, _c)) / _n; _bias -= learning_rate * mlpp_cost.hinge_loss_derivwv(y_hat, current_output_batch_entry, _c)->sum_elements() / _n;
forward_pass(); forward_pass();
@ -305,27 +302,24 @@ real_t MLPPSVC::cost(const Ref<MLPPVector> &z, const Ref<MLPPVector> &y, const R
} }
Ref<MLPPVector> MLPPSVC::evaluatem(const Ref<MLPPMatrix> &X) { Ref<MLPPVector> MLPPSVC::evaluatem(const Ref<MLPPMatrix> &X) {
MLPPLinAlg alg;
MLPPActivation avn; MLPPActivation avn;
return avn.sign_normv(alg.scalar_addnv(_bias, alg.mat_vec_multnv(X, _weights)));
return avn.sign_normv(X->mult_vec(_weights)->scalar_addn(_bias));
} }
Ref<MLPPVector> MLPPSVC::propagatem(const Ref<MLPPMatrix> &X) { Ref<MLPPVector> MLPPSVC::propagatem(const Ref<MLPPMatrix> &X) {
MLPPLinAlg alg; return X->mult_vec(_weights)->scalar_addn(_bias);
MLPPActivation avn;
return alg.scalar_addnv(_bias, alg.mat_vec_multnv(X, _weights));
} }
real_t MLPPSVC::evaluatev(const Ref<MLPPVector> &x) { real_t MLPPSVC::evaluatev(const Ref<MLPPVector> &x) {
MLPPLinAlg alg;
MLPPActivation avn; MLPPActivation avn;
return avn.sign_normr(alg.dotnv(_weights, x) + _bias);
return avn.sign_normr(_weights->dot(x) + _bias);
} }
real_t MLPPSVC::propagatev(const Ref<MLPPVector> &x) { real_t MLPPSVC::propagatev(const Ref<MLPPVector> &x) {
MLPPLinAlg alg;
MLPPActivation avn; MLPPActivation avn;
return alg.dotnv(_weights, x) + _bias; return _weights->dot(x) + _bias;
} }
// sign ( wTx + b ) // sign ( wTx + b )