Reviewed Changes

src/mlpack/tests/main_tests/decision_stump_test.cpp

@@ -112,20 +112,11 @@ BOOST_AUTO_TEST_CASE(DecisionStumpLabelsLessDimensionTest)
 
   size_t testSize = testData.n_cols;
 
-  // Delete the last row containing labels from input dataset
-  // and store it as a new dataset to be used while training
-  // second model.
-  arma::mat inputData2 = inputData;
-  inputData2.shed_row(inputData2.n_rows - 1);
-
-  // Create a copy of testData to be reused.
-  arma::mat testData2 = testData;
-
   // Input training data.
-  SetInputParam("training", std::move(inputData));
+  SetInputParam("training", inputData);
 
   // Input test data.
-  SetInputParam("test", std::move(testData));
+  SetInputParam("test", testData);
 
   mlpackMain();
 
@@ -147,9 +138,12 @@ BOOST_AUTO_TEST_CASE(DecisionStumpLabelsLessDimensionTest)
 
   // Now train DS with labels provided.
 
+  // Delete last row of inputData.
+  inputData.shed_row(inputData.n_rows - 1);
+
   // Input training data.
-  SetInputParam("training", std::move(inputData2));
-  SetInputParam("test", std::move(testData2));
+  SetInputParam("training", std::move(inputData));
+  SetInputParam("test", std::move(testData));
   // Pass Labels.
   SetInputParam("labels", std::move(labels));
 
@@ -186,14 +180,11 @@ BOOST_AUTO_TEST_CASE(DecisionStumpModelReuseTest)
 
   size_t testSize = testData.n_cols;
 
-  // Create a copy of testData to be reused.
-  arma::mat testData2 = testData;
-
   // Input training data.
   SetInputParam("training", std::move(inputData));
 
   // Input test data.
-  SetInputParam("test", std::move(testData));
+  SetInputParam("test", testData);
 
   mlpackMain();
 
@@ -205,7 +196,7 @@ BOOST_AUTO_TEST_CASE(DecisionStumpModelReuseTest)
   CLI::GetSingleton().Parameters()["test"].wasPassed = false;
 
   // Input trained model.
-  SetInputParam("test", std::move(testData2));
+  SetInputParam("test", std::move(testData));
   SetInputParam("input_model",
                 std::move(CLI::GetParam<DSModel>("output_model")));
 

src/mlpack/tests/main_tests/decision_tree_test.cpp

@@ -139,7 +139,7 @@ BOOST_AUTO_TEST_CASE(DecisionModelReuseTest)
   SetInputParam("weights", std::move(weights));
 
   // Input test data.
-  SetInputParam("test", std::move(testData));
+  SetInputParam("test", testData);
 
   mlpackMain();
 
@@ -154,9 +154,6 @@ BOOST_AUTO_TEST_CASE(DecisionModelReuseTest)
   CLI::GetSingleton().Parameters()["weights"].wasPassed = false;
   CLI::GetSingleton().Parameters()["test"].wasPassed = false;
 
-  if (!data::Load("vc2_test.csv", testData))
-    BOOST_FAIL("Cannot load test dataset vc2.csv!");
-
   // Input trained model.
   SetInputParam("test", std::move(testData));
   SetInputParam("input_model",

src/mlpack/tests/main_tests/nbc_test.cpp

@@ -114,20 +114,11 @@ BOOST_AUTO_TEST_CASE(NBCLabelsLessDimensionTest)
 
   size_t testSize = testData.n_cols;
 
-  // Delete the last row containing labels from input dataset
-  // and store it as a new dataset to be used while training
-  // second model.
-  arma::mat inputData2 = inputData;
-  inputData2.shed_row(inputData2.n_rows - 1);
-
-  // Create a copy of testData to be reused.
-  arma::mat testData2 = testData;
-
   // Input training data.
-  SetInputParam("training", std::move(inputData));
+  SetInputParam("training", inputData);
 
   // Input test data.
-  SetInputParam("test", std::move(testData));
+  SetInputParam("test", testData);
 
   mlpackMain();
 
@@ -153,9 +144,11 @@ BOOST_AUTO_TEST_CASE(NBCLabelsLessDimensionTest)
 
   // Now train NBC with labels provided.
 
+  inputData.shed_row(inputData.n_rows - 1);
+
   // Input training data.
-  SetInputParam("training", std::move(inputData2));
-  SetInputParam("test", std::move(testData2));
+  SetInputParam("training", std::move(inputData));
+  SetInputParam("test", std::move(testData));
   // Pass Labels.
   SetInputParam("labels", std::move(labels));
 
@@ -195,14 +188,11 @@ BOOST_AUTO_TEST_CASE(NBCModelReuseTest)
 
   size_t testSize = testData.n_cols;
 
-  // Create a copy of testData to be reused.
-  arma::mat testData2 = testData;
-
   // Input training data.
   SetInputParam("training", std::move(inputData));
 
   // Input test data.
-  SetInputParam("test", std::move(testData));
+  SetInputParam("test", testData);
 
   mlpackMain();
 
@@ -216,7 +206,7 @@ BOOST_AUTO_TEST_CASE(NBCModelReuseTest)
   CLI::GetSingleton().Parameters()["test"].wasPassed = false;
 
   // Input trained model.
-  SetInputParam("test", std::move(testData2));
+  SetInputParam("test", std::move(testData));
   SetInputParam("input_model",
                 std::move(CLI::GetParam<NBCModel>("output_model")));
 
@@ -281,17 +271,11 @@ BOOST_AUTO_TEST_CASE(NBCIncrementalVarianceTest)
 
   size_t testSize = testData.n_cols;
 
-  // Create a copy of inputData to be reused.
-  arma::mat inputData2 = inputData;
-
-  // Create a copy of testData to be reused.
-  arma::mat testData2 = testData;
-
   // Input training data.
-  SetInputParam("training", std::move(inputData));
+  SetInputParam("training", inputData);
 
   // Input test data.
-  SetInputParam("test", std::move(testData));
+  SetInputParam("test", testData);
   SetInputParam("incremental_variance", (bool) true);
 
   mlpackMain();
@@ -320,8 +304,8 @@ BOOST_AUTO_TEST_CASE(NBCIncrementalVarianceTest)
   // Now train NBC without incremental_variance.
 
   // Input training data.
-  SetInputParam("training", std::move(inputData2));
-  SetInputParam("test", std::move(testData2));
+  SetInputParam("training", std::move(inputData));
+  SetInputParam("test", std::move(testData));
   SetInputParam("incremental_variance", (bool) false);
 
   mlpackMain();

src/mlpack/tests/main_tests/perceptron_test.cpp

@@ -112,20 +112,11 @@ BOOST_AUTO_TEST_CASE(PerceptronLabelsLessDimensionTest)
 
   size_t testSize = testData.n_cols;
 
-  // Delete the last row containing labels from input dataset
-  // and store it as a new dataset to be used while training
-  // second model.
-  arma::mat inputData2 = inputData;
-  inputData2.shed_row(inputData2.n_rows - 1);
-
-  // Create a copy of testData to be reused.
-  arma::mat testData2 = testData;
-
   // Input training data.
-  SetInputParam("training", std::move(inputData));
+  SetInputParam("training", inputData);
 
   // Input test data.
-  SetInputParam("test", std::move(testData));
+  SetInputParam("test", testData);
 
   mlpackMain();
 
@@ -140,6 +131,8 @@ BOOST_AUTO_TEST_CASE(PerceptronLabelsLessDimensionTest)
   CLI::GetSingleton().Parameters()["training"].wasPassed = false;
   CLI::GetSingleton().Parameters()["test"].wasPassed = false;
 
+  inputData.shed_row(inputData.n_rows - 1);
+
   // Store outputs.
   arma::Row<size_t> output;
   output = std::move(CLI::GetParam<arma::Row<size_t>>("output"));
@@ -147,8 +140,8 @@ BOOST_AUTO_TEST_CASE(PerceptronLabelsLessDimensionTest)
   // Now train pereptron with labels provided.
 
   // Input training data.
-  SetInputParam("training", std::move(inputData2));
-  SetInputParam("test", std::move(testData2));
+  SetInputParam("training", std::move(inputData));
+  SetInputParam("test", std::move(testData));
   // Pass Labels.
   SetInputParam("labels", std::move(labels));
 
@@ -184,14 +177,11 @@ BOOST_AUTO_TEST_CASE(PerceptronModelReuseTest)
 
   size_t testSize = testData.n_cols;
 
-  // Create a copy of testData to be reused.
-  arma::mat testData2 = testData;
-
   // Input training data.
   SetInputParam("training", std::move(inputData));
 
   // Input test data.
-  SetInputParam("test", std::move(testData));
+  SetInputParam("test", testData);
 
   mlpackMain();
 
@@ -203,7 +193,7 @@ BOOST_AUTO_TEST_CASE(PerceptronModelReuseTest)
   CLI::GetSingleton().Parameters()["test"].wasPassed = false;
 
   // Input trained model.
-  SetInputParam("test", std::move(testData2));
+  SetInputParam("test", std::move(testData));
   SetInputParam("input_model",
                 std::move(CLI::GetParam<PerceptronModel>("output_model")));
 

src/mlpack/tests/main_tests/random_forest_test.cpp

@@ -102,15 +102,12 @@ BOOST_AUTO_TEST_CASE(RandomForestModelReuseTest)
 
   size_t testSize = testData.n_cols;
 
-  // Create a copy of testData to be reused.
-  arma::mat testData2 = testData;
-
   // Input training data.
   SetInputParam("training", std::move(inputData));
   SetInputParam("labels", std::move(labels));
 
   // Input test data.
-  SetInputParam("test", std::move(testData));
+  SetInputParam("test", testData);
 
   mlpackMain();
 
@@ -125,7 +122,7 @@ BOOST_AUTO_TEST_CASE(RandomForestModelReuseTest)
   CLI::GetSingleton().Parameters()["test"].wasPassed = false;
 
   // Input trained model.
-  SetInputParam("test", std::move(testData2));
+  SetInputParam("test", std::move(testData));
   SetInputParam("input_model",
                 std::move(CLI::GetParam<RandomForestModel>("output_model")));
 
@@ -230,34 +227,26 @@ BOOST_AUTO_TEST_CASE(RandomForestDiffMinLeafSizeTest)
   if (!data::Load("vc2_labels.txt", labels))
     BOOST_FAIL("Cannot load labels for vc2_labels.txt");
 
-  // Create copy of training data.
-  arma::mat inputData2 = inputData;
-  arma::Row<size_t> labels2 = labels;
-
   // Input training data.
-  SetInputParam("training", std::move(inputData));
-  SetInputParam("labels", std::move(labels));
+  SetInputParam("training", inputData);
+  SetInputParam("labels", labels);
   SetInputParam("minimum_leaf_size", (int) 20);
 
   mlpackMain();
 
   // Calculate training accuracy.
   arma::Row<size_t> predictions;
-  CLI::GetParam<RandomForestModel>("output_model").rf.Classify(inputData2,
+  CLI::GetParam<RandomForestModel>("output_model").rf.Classify(inputData,
        predictions);
 
-  size_t correct = arma::accu(predictions == labels2);
-  double accuracy20 = (double(correct) / double(labels2.n_elem) * 100);
-
-  // Create copy of training data.
-  inputData = inputData2;
-  labels = labels2;
+  size_t correct = arma::accu(predictions == labels);
+  double accuracy20 = (double(correct) / double(labels.n_elem) * 100);
 
   // Train for minimium leaf size 10.
 
   // Input training data.
-  SetInputParam("training", std::move(inputData2));
-  SetInputParam("labels", std::move(labels2));
+  SetInputParam("training", inputData);
+  SetInputParam("labels", labels);
   SetInputParam("minimum_leaf_size", (int) 10);
 
   mlpackMain();
@@ -269,25 +258,21 @@ BOOST_AUTO_TEST_CASE(RandomForestDiffMinLeafSizeTest)
   correct = arma::accu(predictions == labels);
   double accuracy10 = (double(correct) / double(labels.n_elem) * 100);
 
-  // Create copy of training data.
-  inputData2 = inputData;
-  labels2 = labels;
-
   // Train for minimium leaf size 1.
 
   // Input training data.
-  SetInputParam("training", std::move(inputData));
-  SetInputParam("labels", std::move(labels));
+  SetInputParam("training", inputData);
+  SetInputParam("labels", labels);
   SetInputParam("minimum_leaf_size", (int) 1);
 
   mlpackMain();
 
   // Calculate training accuracy.
-  CLI::GetParam<RandomForestModel>("output_model").rf.Classify(inputData2,
+  CLI::GetParam<RandomForestModel>("output_model").rf.Classify(inputData,
        predictions);
 
-  correct = arma::accu(predictions == labels2);
-  double accuracy1 = (double(correct) / double(labels2.n_elem) * 100);
+  correct = arma::accu(predictions == labels);
+  double accuracy1 = (double(correct) / double(labels.n_elem) * 100);
 
   BOOST_REQUIRE(accuracy1 > accuracy10 && accuracy10 > accuracy20);
 }
@@ -314,12 +299,9 @@ BOOST_AUTO_TEST_CASE(RandomForestDiffNumTreeTest)
   if (!data::Load("vc2_test_labels.txt", testLabels))
     BOOST_FAIL("Cannot load labels for vc2__test_labels.txt");
 
-  // Create copy of training data.
-  arma::mat inputData2 = inputData;
-  arma::Row<size_t> labels2 = labels;
   // Input training data.
-  SetInputParam("training", std::move(inputData));
-  SetInputParam("labels", std::move(labels));
+  SetInputParam("training", inputData);
+  SetInputParam("labels", labels);
   SetInputParam("num_trees", (int) 1);
 
   mlpackMain();
@@ -332,15 +314,11 @@ BOOST_AUTO_TEST_CASE(RandomForestDiffNumTreeTest)
   size_t correct = arma::accu(predictions == testLabels);
   double accuracy1 = (double(correct) / double(testLabels.n_elem) * 100);
 
-  // Create copy of training data.
-  inputData = inputData2;
-  labels = labels2;
-
   // Train for num_trees 5.
 
   // Input training data.
-  SetInputParam("training", std::move(inputData2));
-  SetInputParam("labels", std::move(labels2));
+  SetInputParam("training", inputData);
+  SetInputParam("labels", labels);
   SetInputParam("num_trees", (int) 5);
 
   mlpackMain();
@@ -352,10 +330,6 @@ BOOST_AUTO_TEST_CASE(RandomForestDiffNumTreeTest)
   correct = arma::accu(predictions == testLabels);
   double accuracy5 = (double(correct) / double(testLabels.n_elem) * 100);
 
-  // Create copy of training data.
-  inputData2 = inputData;
-  labels2 = labels;
-
   // Train for num_trees 10.
 
   // Input training data.

src/mlpack/tests/main_tests/softmax_regression_test.cpp

@@ -130,15 +130,12 @@ BOOST_AUTO_TEST_CASE(SoftmaxRegressionModelReuseTest)
 
   size_t testSize = testData.n_cols;
 
-  // Create a copy of testData to be reused.
-  arma::mat testData2 = testData;
-
   // Input training data.
   SetInputParam("training", std::move(inputData));
   SetInputParam("labels", std::move(labels));
 
   // Input test data.
-  SetInputParam("test", std::move(testData));
+  SetInputParam("test", testData);
 
   mlpackMain();
 
@@ -151,7 +148,7 @@ BOOST_AUTO_TEST_CASE(SoftmaxRegressionModelReuseTest)
   CLI::GetSingleton().Parameters()["test"].wasPassed = false;
 
   // Input trained model.
-  SetInputParam("test", std::move(testData2));
+  SetInputParam("test", std::move(testData));
   SetInputParam("input_model",
                 std::move(CLI::GetParam<SoftmaxRegression>("output_model")));
 
@@ -311,18 +308,13 @@ BOOST_AUTO_TEST_CASE(SoftmaxRegressionDiffLambdaTest)
 
   size_t testSize = testData.n_cols;
 
-  // Create a copy of data to be reused.
-  arma::mat inputData2 = inputData;
-  arma::mat testData2 = testData;
-  arma::Row<size_t> labels2 = labels;
-
   // Input training data.
-  SetInputParam("training", std::move(inputData));
-  SetInputParam("labels", std::move(labels));
+  SetInputParam("training", inputData);
+  SetInputParam("labels", labels);
   SetInputParam("lambda", (double) 0.1);
 
   // Input test data.
-  SetInputParam("test", std::move(testData));
+  SetInputParam("test", testData);
 
   mlpackMain();
 
@@ -333,33 +325,25 @@ BOOST_AUTO_TEST_CASE(SoftmaxRegressionDiffLambdaTest)
   // Reset passed parameters.
   CLI::GetSingleton().Parameters()["training"].wasPassed = false;
   CLI::GetSingleton().Parameters()["labels"].wasPassed = false;
-  CLI::GetSingleton().Parameters()["lambda"].wasPassed = false;
   CLI::GetSingleton().Parameters()["test"].wasPassed = false;
 
   // Train SR for lamda 0.9.
 
   // Input training data.
-  SetInputParam("training", std::move(inputData2));
-  SetInputParam("labels", std::move(labels2));
+  SetInputParam("training", std::move(inputData));
+  SetInputParam("labels", std::move(labels));
   SetInputParam("lambda", (double) 0.9);
-  SetInputParam("test", std::move(testData2));
+  SetInputParam("test", std::move(testData));
 
   mlpackMain();
 
   // Check that initial parameters and final parameters matrix
   // using saved model are different.
-  bool flag = false;
-  bool* flagPtr = &flag;
   for (size_t i = 0; i < modelParam.n_elem; ++i)
   {
-    if ((int) (modelParam[i] * 1e+6) != (int) (CLI::GetParam<SoftmaxRegression>
-                                      ("output_model").Parameters()[i] * 1e+6))
-    {
-      *flagPtr = true;
-      break;
-    }
+    BOOST_REQUIRE_NE(modelParam[i],
+        CLI::GetParam<SoftmaxRegression>("output_model").Parameters()[i]);
   }
-  BOOST_REQUIRE_EQUAL(flag, true);
 }
 
 /**
@@ -390,18 +374,13 @@ BOOST_AUTO_TEST_CASE(SoftmaxRegressionDiffMaxItrTest)
 
   size_t testSize = testData.n_cols;
 
-  // Create a copy of data to be reused.
-  arma::mat inputData2 = inputData;
-  arma::mat testData2 = testData;
-  arma::Row<size_t> labels2 = labels;
-
   // Input training data.
-  SetInputParam("training", std::move(inputData));
-  SetInputParam("labels", std::move(labels));
+  SetInputParam("training", inputData);
+  SetInputParam("labels", labels);
   SetInputParam("max_iterations", (int) 500);
 
   // Input test data.
-  SetInputParam("test", std::move(testData));
+  SetInputParam("test", testData);
 
   mlpackMain();
 
@@ -412,33 +391,25 @@ BOOST_AUTO_TEST_CASE(SoftmaxRegressionDiffMaxItrTest)
   // Reset passed parameters.
   CLI::GetSingleton().Parameters()["training"].wasPassed = false;
   CLI::GetSingleton().Parameters()["labels"].wasPassed = false;
-  CLI::GetSingleton().Parameters()["max_iterations"].wasPassed = false;
   CLI::GetSingleton().Parameters()["test"].wasPassed = false;
 
   // Train SR for lamda 0.9.
 
   // Input training data.
-  SetInputParam("training", std::move(inputData2));
-  SetInputParam("labels", std::move(labels2));
+  SetInputParam("training", std::move(inputData));
+  SetInputParam("labels", std::move(labels));
   SetInputParam("max_iterations", (int) 1000);
-  SetInputParam("test", std::move(testData2));
+  SetInputParam("test", std::move(testData));
 
   mlpackMain();
 
   // Check that initial parameters and final parameters matrix
   // using saved model are different.
-  bool flag = false;
-  bool* flagPtr = &flag;
   for (size_t i = 0; i < modelParam.n_elem; ++i)
   {
-    if ((int) (modelParam[i] * 1e+6) != (int) (CLI::GetParam<SoftmaxRegression>
-                                      ("output_model").Parameters()[i] * 1e+6))
-    {
-      *flagPtr = true;
-      break;
-    }
+    BOOST_REQUIRE_NE(modelParam[i],
+        CLI::GetParam<SoftmaxRegression>("output_model").Parameters()[i]);
   }
-  BOOST_REQUIRE_EQUAL(flag, true);
 }
 
 /**
@@ -469,18 +440,13 @@ BOOST_AUTO_TEST_CASE(SoftmaxRegressionDiffInterceptTest)
 
   size_t testSize = testData.n_cols;
 
-  // Create a copy of data to be reused.
-  arma::mat inputData2 = inputData;
-  arma::mat testData2 = testData;
-  arma::Row<size_t> labels2 = labels;
-
   // Input training data.
-  SetInputParam("training", std::move(inputData));
-  SetInputParam("labels", std::move(labels));
+  SetInputParam("training", inputData);
+  SetInputParam("labels", labels);
   SetInputParam("no_intercept", (bool) true);
 
   // Input test data.
-  SetInputParam("test", std::move(testData));
+  SetInputParam("test", testData);
 
   mlpackMain();
 
@@ -497,17 +463,17 @@ BOOST_AUTO_TEST_CASE(SoftmaxRegressionDiffInterceptTest)
   // Train SR for no_intercept.
 
   // Input training data.
-  SetInputParam("training", std::move(inputData2));
-  SetInputParam("labels", std::move(labels2));
-  SetInputParam("test", std::move(testData2));
+  SetInputParam("training", std::move(inputData));
+  SetInputParam("labels", std::move(labels));
+  SetInputParam("test", std::move(testData));
 
   mlpackMain();
 
   // Check that initial parameters has 1 more parameter than
   // final parameters matrix.
   BOOST_REQUIRE_EQUAL(
-    CLI::GetParam<SoftmaxRegression>("output_model").Parameters().n_cols,
-    modelParam.n_cols + 1);
+      CLI::GetParam<SoftmaxRegression>("output_model").Parameters().n_cols,
+      modelParam.n_cols + 1);
 }
 
 BOOST_AUTO_TEST_SUITE_END();