Fix ECOC memory error with new SGVector.

shogun-toolbox · May 6, 2012 · 786359f · 786359f
1 parent 9a18a64
commit 786359f
Show file tree

Hide file tree

Showing 6 changed files with 119 additions and 8 deletions.
diff --git a/examples/undocumented/libshogun/classifier_multiclass_ecoc.cpp b/examples/undocumented/libshogun/classifier_multiclass_ecoc.cpp
@@ -96,7 +96,6 @@ int main(int argc, char** argv)
 	CMath::display_vector(out_labels.vector, out_labels.vlen);
 
 	// Free resources
-	out_labels.destroy_vector();
 	SG_UNREF(mc_svm);
 	SG_UNREF(svm);
 	SG_UNREF(output);

diff --git a/examples/undocumented/libshogun/classifier_multiclass_ecoc_random.cpp b/examples/undocumented/libshogun/classifier_multiclass_ecoc_random.cpp
@@ -0,0 +1,112 @@
+#include <shogun/features/Labels.h>
+#include <shogun/io/StreamingAsciiFile.h>
+#include <shogun/io/SGIO.h>
+#include <shogun/features/StreamingSimpleFeatures.h>
+#include <shogun/features/SimpleFeatures.h>
+#include <shogun/multiclass/ecoc/ECOCStrategy.h>
+#include <shogun/multiclass/ecoc/ECOCRandomSparseEncoder.h>
+#include <shogun/multiclass/ecoc/ECOCHDDecoder.h>
+#include <shogun/machine/LinearMulticlassMachine.h>
+#include <shogun/classifier/svm/LibLinear.h>
+#include <shogun/base/init.h>
+
+#define  EPSILON  1e-5
+
+using namespace shogun;
+
+int main(int argc, char** argv)
+{
+	int32_t num_vectors = 0;
+	int32_t num_feats   = 2;
+
+	init_shogun_with_defaults();
+
+	// Prepare to read a file for the training data
+	char fname_feats[]  = "../data/fm_train_real.dat";
+	char fname_labels[] = "../data/label_train_multiclass.dat";
+	CStreamingAsciiFile* ffeats_train  = new CStreamingAsciiFile(fname_feats);
+	CStreamingAsciiFile* flabels_train = new CStreamingAsciiFile(fname_labels);
+	SG_REF(ffeats_train);
+	SG_REF(flabels_train);
+
+	CStreamingSimpleFeatures< float64_t >* stream_features =
+		new CStreamingSimpleFeatures< float64_t >(ffeats_train, false, 1024);
+
+	CStreamingSimpleFeatures< float64_t >* stream_labels =
+		new CStreamingSimpleFeatures< float64_t >(flabels_train, true, 1024);
+
+	SG_REF(stream_features);
+	SG_REF(stream_labels);
+
+	// Create a matrix with enough space to read all the feature vectors
+	SGMatrix< float64_t > mat = SGMatrix< float64_t >(num_feats, 1000);
+
+	// Read the values from the file and store them in mat
+	SGVector< float64_t > vec;
+	stream_features->start_parser();
+	while ( stream_features->get_next_example() )
+	{
+		vec = stream_features->get_vector();
+
+		for ( int32_t i = 0 ; i < num_feats ; ++i )
+			mat[num_vectors*num_feats + i] = vec[i];
+
+		num_vectors++;
+		stream_features->release_example();
+	}
+	stream_features->end_parser();
+
+	// Create features with the useful values from mat
+	CSimpleFeatures< float64_t >* features = new CSimpleFeatures< float64_t >(mat.matrix, num_feats, num_vectors);
+
+	CLabels* labels = new CLabels(num_vectors);
+	SG_REF(features);
+	SG_REF(labels);
+
+	// Read the labels from the file
+	int32_t idx = 0;
+	stream_labels->start_parser();
+	while ( stream_labels->get_next_example() )
+	{
+		labels->set_int_label( idx++, (int32_t)stream_labels->get_label() );
+		stream_labels->release_example();
+	}
+	stream_labels->end_parser();
+
+	// Create liblinear svm classifier with L2-regularized L2-loss
+	CLibLinear* svm = new CLibLinear(L2R_L2LOSS_SVC);
+	SG_REF(svm);
+
+	// Add some configuration to the svm
+	svm->set_epsilon(EPSILON);
+	svm->set_bias_enabled(true);
+
+	// Create a multiclass svm classifier that consists of several of the previous one
+	CLinearMulticlassMachine* mc_svm = new CLinearMulticlassMachine(
+			new CECOCStrategy(new CECOCRandomSparseEncoder(), new CECOCHDDecoder()), (CDotFeatures*) features, svm, labels);
+	SG_REF(mc_svm);
+
+	// Train the multiclass machine using the data passed in the constructor
+	mc_svm->train();
+
+	// Classify the training examples and show the results
+	CLabels* output = mc_svm->apply();
+
+	SGVector< int32_t > out_labels = output->get_int_labels();
+	CMath::display_vector(out_labels.vector, out_labels.vlen);
+
+	// Free resources
+	SG_UNREF(mc_svm);
+	SG_UNREF(svm);
+	SG_UNREF(output);
+	SG_UNREF(features);
+	SG_UNREF(labels);
+	//SG_UNREF(ffeats_train);
+	//SG_UNREF(flabels_train);
+	SG_UNREF(stream_features);
+	SG_UNREF(stream_labels);
+	exit_shogun();
+
+	return 0;
+}
+
diff --git a/src/shogun/multiclass/ecoc/ECOCOVOEncoder.cpp b/src/shogun/multiclass/ecoc/ECOCOVOEncoder.cpp
@@ -14,7 +14,7 @@ using namespace shogun;
 
 SGMatrix<int32_t> CECOCOVOEncoder::create_codebook(int32_t num_classes)
 {
-    SGMatrix<int32_t> code_book(num_classes*(num_classes-1)/2, num_classes);
+    SGMatrix<int32_t> code_book(num_classes*(num_classes-1)/2, num_classes, true);
     code_book.zero();
     int32_t k=0;
     for (int32_t i=0; i < num_classes; ++i)

diff --git a/src/shogun/multiclass/ecoc/ECOCOVREncoder.cpp b/src/shogun/multiclass/ecoc/ECOCOVREncoder.cpp
@@ -14,7 +14,7 @@ using namespace shogun;
 
 SGMatrix<int32_t> CECOCOVREncoder::create_codebook(int32_t num_classes)
 {
-    SGMatrix<int32_t> code_book(num_classes, num_classes);
+    SGMatrix<int32_t> code_book(num_classes, num_classes, true);
     code_book.set_const(-1);
     for (int32_t i=0; i < num_classes; ++i)
         code_book(i, i) = 1;

diff --git a/src/shogun/multiclass/ecoc/ECOCRandomSparseEncoder.cpp b/src/shogun/multiclass/ecoc/ECOCRandomSparseEncoder.cpp
@@ -51,7 +51,7 @@ SGMatrix<int32_t> CECOCRandomSparseEncoder::create_codebook(int32_t num_classes)
     if (codelen <= 0)
         codelen = get_default_code_length(num_classes);
 
-    SGMatrix<int32_t> best_codebook(codelen, num_classes);
+    SGMatrix<int32_t> best_codebook(codelen, num_classes, true);
     int32_t best_dist = 0;
 
     SGMatrix<int32_t> codebook(codelen, num_classes);

diff --git a/src/shogun/multiclass/ecoc/ECOCStrategy.cpp b/src/shogun/multiclass/ecoc/ECOCStrategy.cpp
@@ -37,14 +37,14 @@ CECOCStrategy::~CECOCStrategy()
     SG_UNREF(m_encoder);
     SG_UNREF(m_decoder);
 
-    m_codebook.destroy_matrix();
+    m_codebook.free_matrix();
 }
 
 void CECOCStrategy::train_start(CLabels *orig_labels, CLabels *train_labels)
 {
     CMulticlassStrategy::train_start(orig_labels, train_labels);
 
-    m_codebook.destroy_matrix();
+    m_codebook.free_matrix();
     m_codebook = m_encoder->create_codebook(m_num_classes);
 }
 
@@ -55,7 +55,7 @@ bool CECOCStrategy::train_has_more()
 
 SGVector<int32_t> CECOCStrategy::train_prepare_next()
 {
-    SGVector<int32_t> subset(m_orig_labels->get_num_labels());
+    SGVector<int32_t> subset(m_orig_labels->get_num_labels(), false);
     int32_t tot=0;
     for (int32_t i=0; i < m_orig_labels->get_num_labels(); ++i)
     {
@@ -77,7 +77,7 @@ SGVector<int32_t> CECOCStrategy::train_prepare_next()
     }
 
     CMulticlassStrategy::train_prepare_next();
-    return SGVector<int32_t>(subset.vector, tot);
+    return SGVector<int32_t>(subset.vector, tot, true);
 }
 
 int32_t CECOCStrategy::decide_label(SGVector<float64_t> outputs)