Numpy学习

在深度学习中，经常要把输入数据，中间结果，最终结果保存下来，直接用Python中的list类型保存和导入不方便，而用Numpy则方便很多 

import numpy as np
a = [1,2,3]
np.save('np_a',a)  # save a

np_a = np.load('np_a.npy') # load the saved numpy

# create a numpy
a = np.ones([1,3])   # (shape,dtype,order)
a = np.zeros([1,3])
a = np.random.random([2,4])  #

# operations: +,-,*,/
a = np.random.random([2,4])
b = np.ones([2,4])
a = a + b   # a - b, a * b, a / b
a = a*4   # 和标量相乘

# index
a[0:3]
a[:3]


# image, voice, text: expressed as array