Searched refs:svd (Results 1 - 22 of 22) sorted by relevance

/external/eigen/doc/snippets/
H A DJacobiSVD_basic.cpp3 JacobiSVD<MatrixXf> svd(m, ComputeThinU | ComputeThinV);
4 cout << "Its singular values are:" << endl << svd.singularValues() << endl;
5 cout << "Its left singular vectors are the columns of the thin U matrix:" << endl << svd.matrixU() << endl;
6 cout << "Its right singular vectors are the columns of the thin V matrix:" << endl << svd.matrixV() << endl;
9 cout << "A least-squares solution of m*x = rhs is:" << endl << svd.solve(rhs) << endl;
/external/eigen/Eigen/src/SVD/
H A DJacobiSVD.h76 void allocate(const JacobiSVD<MatrixType, FullPivHouseholderQRPreconditioner>& svd) argument
78 if (svd.rows() != m_qr.rows() || svd.cols() != m_qr.cols())
81 ::new (&m_qr) QRType(svd.rows(), svd.cols());
83 if (svd.m_computeFullU) m_workspace.resize(svd.rows());
86 bool run(JacobiSVD<MatrixType, FullPivHouseholderQRPreconditioner>& svd, const MatrixType& matrix) argument
91 svd.m_workMatrix = m_qr.matrixQR().block(0,0,matrix.cols(),matrix.cols()).template triangularView<Upper>();
92 if(svd
122 allocate(const JacobiSVD<MatrixType, FullPivHouseholderQRPreconditioner>& svd) argument
133 run(JacobiSVD<MatrixType, FullPivHouseholderQRPreconditioner>& svd, const MatrixType& matrix) argument
159 allocate(const JacobiSVD<MatrixType, ColPivHouseholderQRPreconditioner>& svd) argument
170 run(JacobiSVD<MatrixType, ColPivHouseholderQRPreconditioner>& svd, const MatrixType& matrix) argument
213 allocate(const JacobiSVD<MatrixType, ColPivHouseholderQRPreconditioner>& svd) argument
225 run(JacobiSVD<MatrixType, ColPivHouseholderQRPreconditioner>& svd, const MatrixType& matrix) argument
258 allocate(const JacobiSVD<MatrixType, HouseholderQRPreconditioner>& svd) argument
269 run(JacobiSVD<MatrixType, HouseholderQRPreconditioner>& svd, const MatrixType& matrix) argument
309 allocate(const JacobiSVD<MatrixType, HouseholderQRPreconditioner>& svd) argument
321 run(JacobiSVD<MatrixType, HouseholderQRPreconditioner>& svd, const MatrixType& matrix) argument
367 run(typename SVD::WorkMatrixType& work_matrix, SVD& svd, Index p, Index q, RealScalar& maxDiagEntry) argument
[all...]
/external/eigen/lapack/
H A Dsvd.cpp56 BDCSVD<PlainMatrixType> svd(mat,option);
58 make_vector(s,diag_size) = svd.singularValues().head(diag_size);
62 matrix(u,*m,*m,*ldu) = svd.matrixU();
63 matrix(vt,*n,*n,*ldvt) = svd.matrixV().adjoint();
67 matrix(u,*m,diag_size,*ldu) = svd.matrixU();
68 matrix(vt,diag_size,*n,*ldvt) = svd.matrixV().adjoint();
72 matrix(a,*m,*n,*lda) = svd.matrixU();
73 matrix(vt,*n,*n,*ldvt) = svd.matrixV().adjoint();
77 matrix(u,*m,*m,*ldu) = svd.matrixU();
78 matrix(a,diag_size,*n,*lda) = svd
[all...]
/external/eigen/test/
H A Dsvd_common.h24 void svd_check_full(const MatrixType& m, const SvdType& svd) argument
41 sigma.diagonal() = svd.singularValues().template cast<Scalar>();
42 MatrixUType u = svd.matrixU();
43 MatrixVType v = svd.matrixV();
69 SvdType svd(m, computationOptions);
71 VERIFY_IS_APPROX(svd.singularValues(), referenceSvd.singularValues());
75 VERIFY( (svd.matrixV().adjoint()*svd.matrixV()).isIdentity(prec) );
76 VERIFY_IS_APPROX( svd.matrixV().leftCols(diagSize) * svd
[all...]
H A Dqr_colpivoting.cpp57 JacobiSVD<MatrixType> svd(matrix, ComputeThinU | ComputeThinV);
58 MatrixType svd_solution = svd.solve(rhs);
89 JacobiSVD<MatrixType> svd(matrix, ComputeFullU | ComputeFullV);
90 Matrix<Scalar, Cols, Cols2> svd_solution = svd.solve(rhs);
/external/eigen/Eigen/src/Geometry/
H A DUmeyama.h131 JacobiSVD<MatrixType> svd(sigma, ComputeFullU | ComputeFullV);
139 if ( svd.matrixU().determinant() * svd.matrixV().determinant() < 0 )
143 Rt.block(0,0,m,m).noalias() = svd.matrixU() * S.asDiagonal() * svd.matrixV().transpose();
148 const Scalar c = Scalar(1)/src_var * svd.singularValues().dot(S);
H A DTransform.h1081 JacobiSVD<LinearMatrixType> svd(linear(), ComputeFullU | ComputeFullV);
1083 Scalar x = (svd.matrixU() * svd.matrixV().adjoint()).determinant(); // so x has absolute value 1
1084 VectorType sv(svd.singularValues());
1086 if(scaling) scaling->lazyAssign(svd.matrixV() * sv.asDiagonal() * svd.matrixV().adjoint());
1089 LinearMatrixType m(svd.matrixU());
1091 rotation->lazyAssign(m * svd.matrixV().adjoint());
1110 JacobiSVD<LinearMatrixType> svd(linear(), ComputeFullU | ComputeFullV);
1112 Scalar x = (svd
[all...]
H A DHyperplane.h109 JacobiSVD<Matrix<Scalar,2,3> > svd(m, ComputeFullV);
110 result.normal() = svd.matrixV().col(2);
H A DQuaternion.h596 JacobiSVD<Matrix<Scalar,2,3> > svd(m, ComputeFullV);
597 Vector3 axis = svd.matrixV().col(2);
/external/tensorflow/tensorflow/core/kernels/
H A Dsvd_op_impl.h88 Eigen::BDCSVD<Matrix> svd(inputs[0], options);
89 outputs->at(0) = svd.singularValues().template cast<Scalar>();
91 outputs->at(1) = svd.matrixU();
92 outputs->at(2) = svd.matrixV();
/external/tensorflow/tensorflow/python/kernel_tests/
H A Dsvd_op_test.py42 # The input to svd should be a tensor of at least rank 2.
46 linalg_ops.svd(scalar)
50 linalg_ops.svd(vector)
60 s1, u1, v1 = linalg_ops.svd(
62 s2, u2, v2 = linalg_ops.svd(
66 s1 = linalg_ops.svd(
68 s2 = linalg_ops.svd(
150 s_tf, u_tf, v_tf = linalg_ops.svd(
158 s_tf = linalg_ops.svd(
166 u_np, s_np, v_np = np.linalg.svd(
[all...]
/external/tensorflow/tensorflow/python/ops/linalg/
H A Dlinalg_impl.py51 svd = linalg_ops.svd variable
H A Dlinear_operator.py474 singular_values = linalg_ops.svd(
/external/tensorflow/tensorflow/tools/compatibility/testdata/
H A Dtest_file_v0_11.py174 tf.svd(mat, False, True).eval(),
175 tf.svd(mat, compute_uv=False, full_matrices=True).eval())
/external/tensorflow/tensorflow/contrib/gan/python/eval/python/
H A Dsliced_wasserstein_impl.py190 sig, u = linalg_ops.svd(array_ops.concat([a, b], 0))[:2]
H A Dclassifier_metrics_impl.py102 s, u, v = linalg_ops.svd(mat)
/external/tensorflow/tensorflow/python/ops/
H A Dlinalg_ops.py377 @tf_export('svd', 'linalg.svd')
378 def svd(tensor, full_matrices=False, compute_uv=True, name=None): function
390 s, u, v = svd(a)
391 s = svd(a, compute_uv=False)
417 Mostly equivalent to numpy.linalg.svd, except that
419 `True`, as opposed to `u`, `s`, `v` for numpy.linalg.svd.
421 numpy.linalg.svd.
422 * tf.linalg.svd uses the standard definition of the SVD
425 columns of `v`. On the other hand, numpy.linalg.svd return
[all...]
H A Dlinalg_grad.py277 # https://j-towns.github.io/papers/svd-derivative.pdf
283 s, u, v = linalg_ops.svd(a, compute_uv=True)
327 "svd gradient is not implemented for abs(m - n) > 1 "
/external/tensorflow/tensorflow/contrib/kfac/python/ops/
H A Dutils.py217 evals, evecs, _ = linalg_ops.svd(mat)
225 evals = math_ops.abs(evals) # Should be equivalent to svd approach.
232 "svd": posdef_eig_svd,
H A Dfisher_factors.py1447 sqrtmu, _, E = linalg_ops.svd(hPsi)
/external/tensorflow/tensorflow/python/keras/_impl/keras/preprocessing/
H A Dimage.py832 _, s, vt = linalg.svd(flat_x / np.sqrt(num_examples))
/external/tensorflow/tensorflow/python/framework/
H A Dfunction_test.py912 s, u, v = linalg_ops.svd(m)

Completed in 598 milliseconds