我家有个程儿的博客

keepdims主要用于保持矩阵的二维特性import numpy as npa = np.array([[1,2],[3,4]])# 按行相加，并且保持其二维特性print(np.sum(a, axis=1, keepdims=True))# 按行相加，不保持其二维特性print(np.sum(a, axis=1))输出array([[3], [7]])array([3, 7])...