pmlpp/mlpp/ann/ann.h

73 lines
3.0 KiB
C
Raw Normal View History

2023-01-24 18:57:18 +01:00
#ifndef MLPP_ANN_H
#define MLPP_ANN_H
//
// ANN.hpp
//
// Created by Marc Melikyan on 11/4/20.
//
2023-01-27 13:01:16 +01:00
#include "core/math/math_defs.h"
#include "../hidden_layer/hidden_layer.h"
#include "../output_layer/output_layer.h"
#include "../hidden_layer/hidden_layer_old.h"
#include "../output_layer/output_layer_old.h"
#include <string>
2023-01-24 19:00:54 +01:00
#include <tuple>
#include <vector>
2023-01-24 19:29:29 +01:00
class MLPPANN {
2023-01-24 19:00:54 +01:00
public:
2023-01-27 13:01:16 +01:00
MLPPANN(std::vector<std::vector<real_t>> inputSet, std::vector<real_t> outputSet);
2023-01-24 19:29:29 +01:00
~MLPPANN();
2023-01-27 13:01:16 +01:00
std::vector<real_t> modelSetTest(std::vector<std::vector<real_t>> X);
real_t modelTest(std::vector<real_t> x);
2023-02-04 13:59:26 +01:00
void gradientDescent(real_t learning_rate, int max_epoch, bool UI = false);
void SGD(real_t learning_rate, int max_epoch, bool UI = false);
void MBGD(real_t learning_rate, int max_epoch, int mini_batch_size, bool UI = false);
void Momentum(real_t learning_rate, int max_epoch, int mini_batch_size, real_t gamma, bool NAG, bool UI = false);
void Adagrad(real_t learning_rate, int max_epoch, int mini_batch_size, real_t e, bool UI = false);
void Adadelta(real_t learning_rate, int max_epoch, int mini_batch_size, real_t b1, real_t e, bool UI = false);
void Adam(real_t learning_rate, int max_epoch, int mini_batch_size, real_t b1, real_t b2, real_t e, bool UI = false);
void Adamax(real_t learning_rate, int max_epoch, int mini_batch_size, real_t b1, real_t b2, real_t e, bool UI = false);
void Nadam(real_t learning_rate, int max_epoch, int mini_batch_size, real_t b1, real_t b2, real_t e, bool UI = false);
void AMSGrad(real_t learning_rate, int max_epoch, int mini_batch_size, real_t b1, real_t b2, real_t e, bool UI = false);
2023-01-27 13:01:16 +01:00
real_t score();
2023-01-24 19:00:54 +01:00
void save(std::string fileName);
2023-01-27 13:01:16 +01:00
void setLearningRateScheduler(std::string type, real_t decayConstant);
void setLearningRateScheduler(std::string type, real_t decayConstant, real_t dropRate);
2023-01-24 19:00:54 +01:00
2023-01-27 13:01:16 +01:00
void addLayer(int n_hidden, std::string activation, std::string weightInit = "Default", std::string reg = "None", real_t lambda = 0.5, real_t alpha = 0.5);
void addOutputLayer(std::string activation, std::string loss, std::string weightInit = "Default", std::string reg = "None", real_t lambda = 0.5, real_t alpha = 0.5);
2023-01-24 19:00:54 +01:00
private:
2023-01-27 13:01:16 +01:00
real_t applyLearningRateScheduler(real_t learningRate, real_t decayConstant, real_t epoch, real_t dropRate);
2023-01-24 19:00:54 +01:00
2023-01-27 13:01:16 +01:00
real_t Cost(std::vector<real_t> y_hat, std::vector<real_t> y);
2023-01-24 19:00:54 +01:00
void forwardPass();
2023-01-27 13:01:16 +01:00
void updateParameters(std::vector<std::vector<std::vector<real_t>>> hiddenLayerUpdations, std::vector<real_t> outputLayerUpdation, real_t learning_rate);
std::tuple<std::vector<std::vector<std::vector<real_t>>>, std::vector<real_t>> computeGradients(std::vector<real_t> y_hat, std::vector<real_t> outputSet);
2023-01-24 19:00:54 +01:00
2023-01-27 13:01:16 +01:00
void UI(int epoch, real_t cost_prev, std::vector<real_t> y_hat, std::vector<real_t> outputSet);
2023-01-24 19:00:54 +01:00
2023-01-27 13:01:16 +01:00
std::vector<std::vector<real_t>> inputSet;
std::vector<real_t> outputSet;
std::vector<real_t> y_hat;
2023-01-24 19:00:54 +01:00
std::vector<MLPPOldHiddenLayer> network;
MLPPOldOutputLayer *outputLayer;
2023-01-24 19:00:54 +01:00
int n;
int k;
std::string lrScheduler;
2023-01-27 13:01:16 +01:00
real_t decayConstant;
real_t dropRate;
2023-01-24 19:00:54 +01:00
};
#endif /* ANN_hpp */