Tensorflow圖級別隨機數設置-tf.set_random_seed(seed)

 

tf.set_random_seed(seed)

可使得所有會話中op產生的隨機序列是相等可重復的。

例如：

tf.set_random_seed(1234)

a = tf.random_uniform([1])

b = tf.random_normal([1])

 

# Repeatedly running this block with the same graph will generate different

# sequences of 'a' and 'b'.

print("Session 1")

with tf.Session() as sess1:

  print(sess1.run(a))  # generates 'A1'

  print(sess1.run(a))  # generates 'A2'

  print(sess1.run(b))  # generates 'B1'

  print(sess1.run(b))  # generates 'B2'

 

print("Session 2")

with tf.Session() as sess2:

  print(sess2.run(a))  # generates 'A1'

  print(sess2.run(a))  # generates 'A2'

  print(sess2.run(b))  # generates 'B1'

  print(sess2.run(b))  # generates 'B2'