# -*- coding: UTF-8 -*- # 引入tensorflow import tensorflow as tf # 构造图(Graph)的结构 # 用一个线性方程的例子 y = W * x + b W = tf.Variable(2.0, dtype=tf.float32, name="Weight") # 权重 b = tf.Variable(1.0, dtype=tf.float32, name="Bias") # 偏差 x = tf.placeholder(dtype=tf.float32, name="Input") # 输入 with tf.name_scope("Output"): # 输出的命名空间 y = W * x + b # 输出 #const = tf.constant(2.0) # 不需要初始化 # 定义保存日志的路径 path = "./log" # 创建用于初始化所有变量(Variable)的操作 init = tf.global_variables_initializer() # 创建Session(会话) with tf.Session() as sess: sess.run(init) # 初始化变量 writer = tf.summary.FileWriter(path, sess.graph) result = sess.run(y, {x: 3.0}) print("y = %s" % result) # 打印 y = W * x + b 的值,就是 7