numpy 多维矩阵

来源：互联网发布：windows dns配置文件编辑：程序博客网时间：2024/06/04 18:41

对于shape是（4，3，3，2）的4维矩阵，其意思是在（4，3）的矩阵中，其每个元素是（3，2）的矩阵，也就是（4，3）的矩阵嵌套一个（3，2）的矩阵

import numpy as npa = np.random.randn(4,3,3,2)print(a)

输出结果为

[[[[-2.26728458 -0.93044922]   [-0.79551665 -0.94800901]   [ 0.78514497  2.06200093]]  [[-1.60940096 -0.24614001]   [ 0.29647288 -0.29924968]   [ 0.16358296  0.74884066]]  [[ 1.09777908  1.49181953]   [ 0.46415243  0.24711349]   [-1.2020132   1.24691265]]] [[[-0.31466526 -0.72276109]   [-0.81892115  1.06496607]   [ 1.401888    0.59864402]]  **[[-2.08980204  0.55446417]   [ 1.41521024 -2.51285901]   [ 0.39136998  1.00890537]]**  [[-0.39626076  1.38580398]   [-1.29230797 -1.69083911]   [ 1.54104362  0.8584694 ]]] [[[-0.28891173 -2.07050353]   [-0.36509318 -1.03264759]   [ 0.96609009 -0.84523115]]  [[-1.67746496  0.0796231 ]   [-0.24992364 -0.49804146]   [-0.79940273  0.88040293]]  [[-1.41201124  0.23929839]   [-0.78763281 -0.76797591]   [-1.15122065  0.40700796]]] [[[-1.19984005 -1.81134743]   [-0.72768683 -0.08521888]   [ 0.5493634   1.275239  ]]  [[-1.59056989 -0.97041589]   [-0.74594324 -0.61091165]   [-1.09071522  1.08896788]]  [[ 0.68662254 -1.70185507]   [-0.39004449 -0.47509404]   [-0.11858611 -0.49051322]]]]

那么

b = a[1,1]
print(b)

输出结果为

[[-2.08980204  0.55446417] [ 1.41521024 -2.51285901] [ 0.39136998  1.00890537]]

b刚好是在（4，3）矩阵的（1，1）位置上（矩阵维数从0开始计）

阅读全文

0 0