Tensorflow | Random |常用函数介绍

根据官网的帮助文档，介绍Random类型的函数，方便自己学习和查看。若是有幸帮到别的朋友，深感荣幸。

rf.random_normal

产生正态随机分布

格式：tf.random_normal(shape,mean=0.0,stddev=1.0,dtype=tf.float32,seed=None,name=None)

shape定义维度，mean定义均值，stddev定义方差，dtype定义类型，seed定义种子，name定义名称

例子：

import tensorflow as tf
# Create a tensor of shape [2, 3] consisting of random normal values, with mean
# -1 and standard deviation 4.
norm = tf.random_normal(shape=[2, 3], mean=-1, stddev=4)
with tf.Session() as sess:
print (sess.run(norm))

结果：

[[ -7.80873823 -10.97159195 -11.99345589]

[ 1.79066849 -4.10513306 4.37571764]]

tf.truncated_normal

产生标准正态分布

格式：tf.truncated_normal(shape, mean=0.0, stddev=1.0, dtype=tf.float32, seed=None, name=None)

shape定义维度，mean定义均值，stddev定义方差，dtype定义类型，seed定义种子，name定义名称

例子：

import tensorflow as tf
# Create a tensor of shape [2, 3] consisting of random normal values, with mean
# 0 and standard deviation 1.
norm = tf.truncated_normal(shape=[2,3],mean=0,stddev=1)
with tf.Session() as sess:
print (sess.run(norm))

结果：

[[ 1.89490759 -1.03072059 0.2172989 ]

[-0.29377019 -0.38990787 -1.09539473]]

tf.random_uniform

产生均匀分布

格式：tf.random_uniform(shape, minval=0.0, maxval=1.0, dtype=tf.float32, seed=None, name=None)

shape定义维度，minval区间最小值，maxval区间最大值，dtype定义类型，seed定义种子，name定义名称

例子：

import tensorflow as tf
# Create a tensor of shape [2, 3] consisting of random uniform values, with minval=1
#  and maxval =3.
norm = tf.random_uniform(shape=[2,3],minval=1,maxval=3)
with tf.Session() as sess:
print (sess.run(norm))

结果：

[[ 2.73986316 1.50323987 1.64366412]

[ 1.12579513 1.52106118 1.29330397]]

-tf.random_shuffle

随机的交换位置

格式：tf.random_shuffle(value, seed=None, name=None)

value是一个给定的张量，seed定义的种子，name定义名称

例子：

import tensorflow as tf
c = tf.constant([[1,2],[3,4],[5,6]])
shuff = tf.random_shuffle(value=c,seed=1,name="shuff")
with tf.Session() as sess:
print (sess.run(shuff))

结果：

[[1 2]

[5 6]

[3 4]]

tf.set_random_seed

设置种子

格式：tf.set_random_seed(seed)

seed是给定的种子

例子：

import tensorflow as tf
tf.set_random_seed(1234)
a = tf.random_uniform([1])
b = tf.random_normal([1])
with tf.Session() as sess:
print (sess.run(a))
print (sess.run(b))

结果：

[ 0.59309709]

[ 0.32048994]

每次运行结果都不一致。要一致还是在定义张量的内部来设置。