int main (int argc, const char * argv[])
{
//Parse the data filename from the argument list
if( argc != 2 ){
cout << "Error: failed to parse data filename from command line. You should run this example with one argument pointing to the data filename!\n";
return EXIT_FAILURE;
}
const string filename = argv[1];
//Create a new Softmax instance
Softmax softmax;
//Load some training data to train the classifier
ClassificationData trainingData;
if( !trainingData.load( filename ) ){
cout << "Failed to load training data: " << filename << endl;
return EXIT_FAILURE;
}
//Use 20% of the training dataset to create a test dataset
ClassificationData testData = trainingData.split( 80 );
//Train the classifier
if( !softmax.train( trainingData ) ){
cout << "Failed to train classifier!\n";
return EXIT_FAILURE;
}
//Save the Softmax model to a file
if( !softmax.save("SoftmaxModel.grt") ){
cout << "Failed to save the classifier model!\n";
return EXIT_FAILURE;
}
//Load the Softmax model from a file
if( !softmax.load("SoftmaxModel.grt") ){
cout << "Failed to load the classifier model!\n";
return EXIT_FAILURE;
}
//Use the test dataset to test the softmax model
double accuracy = 0;
for(UINT i=0; i<testData.getNumSamples(); i++){
//Get the i'th test sample
UINT classLabel = testData[i].getClassLabel();
VectorFloat inputVector = testData[i].getSample();
//Perform a prediction using the classifier
if( !softmax.predict( inputVector ) ){
cout << "Failed to perform prediction for test sample: " << i <<"\n";
return EXIT_FAILURE;
}
//Get the predicted class label
UINT predictedClassLabel = softmax.getPredictedClassLabel();
VectorFloat classLikelihoods = softmax.getClassLikelihoods();
VectorFloat classDistances = softmax.getClassDistances();
//Update the accuracy
if( classLabel == predictedClassLabel ) accuracy++;
cout << "TestSample: " << i << " ClassLabel: " << classLabel << " PredictedClassLabel: " << predictedClassLabel << endl;
}
cout << "Test Accuracy: " << accuracy/double(testData.getNumSamples())*100.0 << "%" << endl;
return EXIT_SUCCESS;
}
int main (int argc, const char * argv[])
{
//Create a new Softmax instance
Softmax softmax;
//Load some training data to train the classifier
ClassificationData trainingData;
if( !trainingData.loadDatasetFromFile("SoftmaxTrainingData.txt") ){
cout << "Failed to load training data!\n";
return EXIT_FAILURE;
}
//Use 20% of the training dataset to create a test dataset
ClassificationData testData = trainingData.partition( 80 );
//Train the classifier
if( !softmax.train( trainingData ) ){
cout << "Failed to train classifier!\n";
return EXIT_FAILURE;
}
//Save the Softmax model to a file
if( !softmax.saveModelToFile("SoftmaxModel.txt") ){
cout << "Failed to save the classifier model!\n";
return EXIT_FAILURE;
}
//Load the Softmax model from a file
if( !softmax.loadModelFromFile("SoftmaxModel.txt") ){
cout << "Failed to load the classifier model!\n";
return EXIT_FAILURE;
}
//Use the test dataset to test the softmax model
double accuracy = 0;
for(UINT i=0; i<testData.getNumSamples(); i++){
//Get the i'th test sample
UINT classLabel = testData[i].getClassLabel();
vector< double > inputVector = testData[i].getSample();
//Perform a prediction using the classifier
if( !softmax.predict( inputVector ) ){
cout << "Failed to perform prediction for test sample: " << i <<"\n";
return EXIT_FAILURE;
}
//Get the predicted class label
UINT predictedClassLabel = softmax.getPredictedClassLabel();
vector< double > classLikelihoods = softmax.getClassLikelihoods();
vector< double > classDistances = softmax.getClassDistances();
//Update the accuracy
if( classLabel == predictedClassLabel ) accuracy++;
cout << "TestSample: " << i << " ClassLabel: " << classLabel << " PredictedClassLabel: " << predictedClassLabel << endl;
}
cout << "Test Accuracy: " << accuracy/double(testData.getNumSamples())*100.0 << "%" << endl;
return EXIT_SUCCESS;
}
int main(int argc, char* argv[])
{
Config config;
Layer l(config);
FileReader fr(config.train_file, config.batchSize);
Softmax softmax;
softmax.initSoftmax(config.outputVocabSize, config.hiddenSize, config.batchSize, &l);
l.softmax = &softmax;
type prevPerplexity = 10000000;
if(config.loadWeights)
{
l.loadWeights("Layer_Weights.txt");
softmax.loadWeights("Softmax_Weights.txt");
if(config.verbose)
{
softmax.printMatrices();
l.printMatrices();
}
}
if(config.trainModel)
{
for(int i = 0; i < config.numPasses; i++)
{
std::chrono::time_point<std::chrono::system_clock> start, end, batchStart, batchEnd;
start = std::chrono::system_clock::now();
std::cout << "PASS NUMBER: " << i << std::endl;
fr.resetFile();
for(int j = 0; j < config.numTrainBatches; j++)
{
if( j % 10 == 0){
batchEnd = std::chrono::system_clock::now();
std::chrono::duration<type> elapsed_seconds = batchEnd-batchStart;
printf("Time Taken %f \n", elapsed_seconds.count());
std::cout << "Batch number: " << j << std::endl;
batchStart = std::chrono::system_clock::now();
}
fr.readNextBatch();
l.initLSTMVector(fr);
l.forwardPropagation(fr.inputMiniBatch);
l.backwardPropagation(fr.outputMiniBatch);
if(config.checkGradients){
l.checkGradients(config.ep);
softmax.checkGradients(config.ep);
}
softmax.applyGradients();
l.applyGradients();
l.clearRunningErrors();
l.clearGradients();
softmax.clearGradients();
if( j % 100 == 0 && j != 0){
if( config.getPerplexity)
{
type perplexity = 0.0;
int numWords = 0;
FileReader p_fr(config.test_file, config.batchSize);
for (int i = 0; i < config.numTestBatches; i++){
p_fr.readNextBatch();
l.initLSTMVector(p_fr);
perplexity += l.getError();
l.clearRunningErrors();
l.clearGradients();
softmax.clearGradients();
numWords +=p_fr.numWordsInBatch;
}
perplexity = perplexity/std::log(2.0); //Change to base 2 log
perplexity = std::pow(2,-1*perplexity/numWords);
std::cout << "PERPLEXITY: " << perplexity << std::endl;
}
}
}
end = std::chrono::system_clock::now();
std::chrono::duration<type> elapsed_seconds = end-start;
std::time_t end_time = std::chrono::system_clock::to_time_t(end);
std::cout << "Finished Computing 1 Pass in " << elapsed_seconds.count() << "s\n";
l.dumpMatrices();
softmax.dumpMatrices();
}
if(config.verbose)
{
l.printMatrices();
softmax.printMatrices();
}
}
return 0;
}
