Add paddle.linalg.matrix_power OP

[haolin-nju]

PR types

New features

PR changes

OPs

Describe

This PR supports matrix_power calculation for linear algorithm module of Paddle. One may call paddle.linalg.matrix_power or paddle.matrix_power to use it.

An example is listed below, which is a python3 script:

import paddle

x = paddle.to_tensor([[1, 2, 3],
                      [1, 4, 9],
                      [1, 8, 27]], dtype='float64')
print(paddle.matrix_power(x, 2))
# [[6.  , 34. , 102.],
#  [14. , 90. , 282.],
#  [36. , 250., 804.]]

print(paddle.matrix_power(x, 0))
# [[1., 0., 0.],
#  [0., 1., 0.],
#  [0., 0., 1.]]

print(paddle.matrix_power(x, -2))
# [[ 12.91666667, -12.75000000,  2.83333333 ],
#  [-7.66666667 ,  8.         , -1.83333333 ],
#  [ 1.80555556 , -1.91666667 ,  0.44444444 ]]

A detailed doc PR in Chinese is now available at docs PR3727. Its preview version is available at matrix_power doc preview before UTC+8 20/08/2021.

[paddle-bot-old]

Thanks for your contribution!
Please wait for the result of CI firstly. See Paddle CI Manual for details.

[haolin-nju]

Resolve comments in matrix_power api.

[JiabinYang]

LGTM

[haolin-nju]

Need more advice about naming method and comment message.

[haolin-nju]

Three comments have been resolved in the commit(ce2fa089)

[JiabinYang]

LGTM

[juncaipeng]

LGTM for changing op_threshold_white_list.py

[XiaoguangHu01]

LG API