我对 tensorflow 中的符号样式程序有一个基本问题,下面是代码( tensorflow 0.12 ):
import numpy as np
import tensorflow as tf
import sys
x = tf.placeholder(tf.float32, 1)
y = 1
z = x + y * 2
# x = z
with tf.Session() as sess:
for i in range(2):
result = sess.run([z], feed_dict={x: [1]})
print "result = "
print result[0]
如果注释了'x = z',这是一个典型的有向无环图(DAG),结果是:
result =
[ 3.]
result =
[ 3.]
如果 'x = z' 未注释,则结果为:
result =
[ 1.]
result =
[ 1.]
我认为结果应该与“x = z”语句无关,结果应该是3。是什么导致了差异?谢谢!