lu_unpack

paddle.linalg. lu_unpack ( x, y, unpack_ludata=True, unpack_pivots=True, name=None ) [source]

Unpack L U and P to single matrix tensor . unpack L and U matrix from LU, unpack permutation matrix P from Pivtos .

P mat can be get by pivots:

System Message: ERROR/3 (/usr/local/lib/python3.8/site-packages/paddle/tensor/linalg.py:docstring of paddle.tensor.linalg.lu_unpack, line 6)

Error in “code-block” directive: maximum 1 argument(s) allowed, 15 supplied.

.. code-block:: text
    ones = eye(rows) #eye matrix of rank rows
    for i in range(cols):
        swap(ones[i], ones[pivots[i]])
Parameters

  • x (Tensor) – The LU tensor get from paddle.linalg.lu, which is combined by L and U.

  • y (Tensor) – Pivots get from paddle.linalg.lu.

  • unpack_ludata (bool,optional) – whether to unpack L and U from x. Default: True.

  • unpack_pivots (bool, optional) – whether to unpack permutation matrix P from Pivtos. Default: True.

  • name (str, optional) – Name for the operation (optional, default is None). For more information, please refer to Name.

Returns

P (Tensor), Permutation matrix P of lu factorization.

L (Tensor), The lower triangular matrix tensor of lu factorization.

U (Tensor), The upper triangular matrix tensor of lu factorization.

Examples

import paddle

x = paddle.to_tensor([[1.0, 2.0], [3.0, 4.0], [5.0, 6.0]]).astype('float64')
lu,p,info = paddle.linalg.lu(x, get_infos=True)

# >>> lu:
# Tensor(shape=[3, 2], dtype=float64, place=CUDAPlace(0), stop_gradient=True,
#    [[5.        , 6.        ],
#        [0.20000000, 0.80000000],
#        [0.60000000, 0.50000000]])
# >>> p
# Tensor(shape=[2], dtype=int32, place=CUDAPlace(0), stop_gradient=True,
#    [3, 3])
# >>> info
# Tensor(shape=[], dtype=int32, place=CUDAPlace(0), stop_gradient=True,
#    0)

P,L,U = paddle.linalg.lu_unpack(lu,p)

# >>> P
# (Tensor(shape=[3, 3], dtype=float64, place=CUDAPlace(0), stop_gradient=True,
# [[0., 1., 0.],
# [0., 0., 1.],
# [1., 0., 0.]]),
# >>> L
# Tensor(shape=[3, 2], dtype=float64, place=CUDAPlace(0), stop_gradient=True,
# [[1.        , 0.        ],
# [0.20000000, 1.        ],
# [0.60000000, 0.50000000]]),
# >>> U
# Tensor(shape=[2, 2], dtype=float64, place=CUDAPlace(0), stop_gradient=True,
# [[5.        , 6.        ],
# [0.        , 0.80000000]]))

# one can verify : X = P @ L @ U ;