Julien Schueller
Julien Schueller
... is defined for sample but not for distribution which one is right ?
``` import openturns as ot coll = ot.DistributionFactory.GetContinuousUniVariateFactories() coll.add(ot.DistributionFactory.GetDiscreteUniVariateFactories()) for factory in coll: distribution = factory.build() for p in [0.0,1.0]: q = distribution.computeQuantile(p)[0] try: sq = distribution.computeScalarQuantile(p) except Exception as...
Seems dpotrf used for the cholesky decomposition gives different results across different platforms Here is a minimal reproducer: ``` TBB::Disable(); PlatformInfo::SetNumericalPrecision(15); Sample input_sample(Normal(2).getSample(6)); MaternModel covariance(Point(2, 1.0), 2.5); CovarianceMatrix C(covariance.discretize(input_sample)); std::cout
on mingw (win32), compiling netcdf with cmake, with a cmake-compiled hdf5 using its exported target (module mode, not config mode), results in an installed netcdf-targets.cmake that contains the hdf5::hdf5 target...
I got compilation error due to the openmp syntax used in base/tools/sle/solver/Gmmpp.cpp,Armadillo.cpp,Eigen.cpp with gcc 9.x look for lines containing "#pragma omp parallel if (system.isCloneable()) shared(system, A, nnz, rowsDone)" a possible...
dgelsx, dgegv, dgeqpf are deprecated since lapack 3.1.1 (2007), remove them
the wingweight example that use HSIC takes much longer to run when the doe size is set from 200 to 800 (see #2150) is there something that explains the computing...
Here's my original patch to drop small values in pie chart: ``` diff --git a/lib/etc/openturns.conf.in b/lib/etc/openturns.conf.in index c9b0d3d9b..056f13195 100644 --- a/lib/etc/openturns.conf.in +++ b/lib/etc/openturns.conf.in @@ -79,6 +79,7 @@ + diff --git...