Trainer Class Reference

Class that makes training a network easier. It uses a special stop strategy object to decide when training is complete. Trainer class creates a network configuration that is optimal according to the used stop training strategy. More...

#include <Trainer.h>

Collaboration diagram for Trainer:

List of all members.

Classes
struct	Stats
	Basic statistics provided by neural network trainer. More...
Public Member Functions
	Trainer (Neural_Net *net, const std::vector< double > *training_set)
	Constructor.
void	set_net (Neural_Net *net)
	Sets a Neural_Net object to train.
void	set_training_set (const std::vector< double > *training_set)
	Sets training data for coaching the net.
void	set_Stop_Training_Strategy (Stop_Training_Strategy *Stop_Training_Strategy)
	Sets Stop_Training_Strategy object that controls when the training is considered complete.
void	set_data_interpreter (Data_Set_Interpreter *interpreter)
	Sets Data_Set_Interpreter object that controls how training set is divided into training samples.
void	set_net_config_output_stream (Simple_Stream< char > *trained_net_config)
	After the training network configuration will be written to the provided stream.
Stats	get_stats ()
	Returns Stats structure filled with current training session statistics.
int	start ()
	Starts network training. Process continues until Stop_Training_Strategy object instructs otherwise.
void	reset_stats ()
	Resets statistics gathered while training.

---

Detailed Description

Class that makes training a network easier. It uses a special stop strategy object to decide when training is complete. Trainer class creates a network configuration that is optimal according to the used stop training strategy.