Tensorflow 随机张量

Tensorflow 随机张量

• tf.random_normal(shape, mean, stddev, dtype, seed, name)

    正太分布：生成100个均值为0、标准差为1的随机数。

import tensorflow as tf
import numpy as np
import matplotlib.pyplot as plt

norm = tf.random_normal([100], mean = 0.0, stddev = 1.0) 
with tf.Session() as sess:
    y = sess.run(norm)
    plt.plot(y)
    plt.show()

 

• tf.truncated_normal(shape, mean, stddev, dtype, seed, name)

正太分布：生成的值遵循具有指定平均值和标准偏差的正态分布，与tf.random_normal()不同之处在于其平均值大于 2 个标准差的值将被丢弃并重新选择。

tru_norm = tf.truncated_normal([100], mean = 0.0, stddev = 1.0) 
with tf.Session() as sess:
    y = sess.run(tru_norm)
    plt.plot(y)
    plt.show()