Hi Marcus, The test data has already been copied to the binaries directory.
There are a number of problems: (1) It seems MSVC-2015 has a problem with parallel_sgd and associated implementations. Under debug the compiler throws an internal compiler error. Under release it compiles fine but fails the unit tests. (2) Any dependencies upon parallel_sgd also fail (regularized_svd) (3) The unit tests have problem loading the braziltourism dataset, despite the dataset file being present. Thanks Jason On Mon, Jun 11, 2018 at 10:04 PM, Marcus Edel <[email protected]> wrote: > Hello Jason, > > Looks like the test datasets aren't accessible from where you start the > test, > one idea is to copy the datasets into C:\projects\mlpack-3.0.2\ > build\Release. > Let me know if that solves the issue. > > Thanks, > Marcus > > > On 10. Jun 2018, at 16:36, Jason Liu <[email protected]> wrote: > > > > Hi, > > > > I built mlpack under Windows 10 using MSVC-2015. I have all dependencies > (boost, armadillo, blas, lapack) built from source using the same compiler. > All went fine without errors. But when I run the unit tests > (mlpack_test.exe) I get the following test failures: > > > > C:\projects\mlpack-3.0.2\build\Release>mlpack_test.exe > > Running 1438 test cases... > > unknown location(0): fatal error: in "LoadSaveTest/CaseTest": class > std::runtime_error: Parse error at line 13 token 0: "TA". > > C:\projects\mlpack-3.0.2\src\mlpack\tests\load_save_test.cpp(1810): > last checkpoint: "CaseTest" entry. > > C:/projects/mlpack-3.0.2/src/mlpack/tests/parallel_sgd_test.cpp(69): > fatal error: in "ParallelSGDTest/SimpleParallelSGDTest": difference{inf} > between result{inf} and 123.75{123.75} exceeds 0.01% > > C:/projects/mlpack-3.0.2/src/mlpack/tests/parallel_sgd_test.cpp(100): > fatal error: in "ParallelSGDTest/GeneralizedRosenbrockTest": absolute > value of result{inf} exceeds 1e-08 > > C:/projects/mlpack-3.0.2/src/mlpack/tests/regularized_svd_test.cpp(316): > fatal error: in "RegularizedSVDTest/RegularizedSVDFunctionOptimizeHOGWILD": > absolute value of relativeError{-nan(ind)} exceeds 0.01 > > C:/projects/mlpack-3.0.2/src/mlpack/tests/main_tests/decision_tree_test.cpp(103): > fatal error: in > "DecisionTreeMainTest/DecisionTreeCategoricalOutputDimensionTest": > Cannot load train dataset braziltourism.arff! > > C:/projects/mlpack-3.0.2/src/mlpack/tests/main_tests/decision_tree_test.cpp(149): > fatal error: in "DecisionTreeMainTest/DecisionTreeMinimumLeafSizeTest": > Cannot load train dataset braziltourism.arff! > > C:/projects/mlpack-3.0.2/src/mlpack/tests/main_tests/decision_tree_test.cpp(178): > fatal error: in "DecisionTreeMainTest/DecisionMinimumGainSplitTest": > Cannot load train dataset braziltourism.arff! > > C:/projects/mlpack-3.0.2/src/mlpack/tests/main_tests/decision_tree_test.cpp(207): > fatal error: in "DecisionTreeMainTest/DecisionRegularisationTest": Cannot > load train dataset braziltourism.arff! > > C:/projects/mlpack-3.0.2/src/mlpack/tests/main_tests/decision_tree_test.cpp(362): > fatal error: in "DecisionTreeMainTest/DecisionModelCategoricalReuseTest": > Cannot load train dataset braziltourism.arff! > > ** On entry to DLASCL, parameter number 4 had an illegal value > > ** On entry to DLASCL, parameter number 5 had an illegal value > > C:/projects/mlpack-3.0.2/src/mlpack/tests/main_tests/hoeffding_tree_test.cpp(99): > fatal error: in > "HoeffdingTreeMainTest/HoeffdingTreeCategoricalOutputDimensionTest": > Cannot load train dataset braziltourism.arff! > > C:/projects/mlpack-3.0.2/src/mlpack/tests/main_tests/hoeffding_tree_test.cpp(295): > fatal error: in "HoeffdingTreeMainTest/HoeffdingModelCategoricalReuseTest": > Cannot load train dataset braziltourism.arff! > > > > *** 11 failures are detected in the test module "mlpackTest" > > > > Does anyone know how to fix these test failures? > > > > Thanks! > > > > Kind Regards > > Jason > > > > _______________________________________________ > > mlpack mailing list > > [email protected] > > http://knife.lugatgt.org/cgi-bin/mailman/listinfo/mlpack > >
_______________________________________________ mlpack mailing list [email protected] http://knife.lugatgt.org/cgi-bin/mailman/listinfo/mlpack
