tf学习笔记 03 tf.Variable

当训练一个模型时，使用Variable保存和更新参数，variable是一块存储的内存，variable需要显示的initialized

而且也可以在训练过程和结束后存在到硬盘中，也可以重新载入到模型

# Create a variable with a random value.
weights = tf.Variable(tf.random_normal([784, 200], stddev=0.35),
                      name="weights")
# Create another variable with the same value as 'weights'.
w2 = tf.Variable(weights.initialized_value(), name="w2")
# Create another variable with twice the value of 'weights'
w_twice = tf.Variable(weights.initialized_value() * 0.2, name="w_twice")

创建：

当你创建一个Variable是，你需要传递一个Tensor给Variable()构造器。通常是常量或者随机值。Variable创建需要

指定的shape，tensorflow提供了修改variable shape的方法

# Create two variables.
weights = tf.Variable(tf.random_normal([784, 200], stddev=0.35),
                      name="weights")
biases = tf.Variable(tf.zeros([200]), name="biases")

调用 tf.variable(）方法添加了以下几个ops到graph中：

1，一个 Varible操作持有Variable变量

2，一个初始化（initializer）操作，将初始值赋值给Variable， 这实际上是一个 tf.assign操作

3，正对初始值得操作，例如，zeros()操作和random_normal()操作添加到graph中

返回一个python类 tf.Variable

初始化：

Variable的初始化必须在运行graph之前显示地运行，最方便的方法时 运行 tf.initialize_all_variables().

通过其他Variable初始化

# Create a variable with a random value.
weights = tf.Variable(tf.random_normal([784, 200], stddev=0.35),
                      name="weights")
# Create another variable with the same value as 'weights'.
w2 = tf.Variable(weights.initialized_value(), name="w2")
# Create another variable with twice the value of 'weights'
w_twice = tf.Variable(weights.initialized_value() * 0.2, name="w_twice")

关于Variable详细信息

http://wiki.jikexueyuan.com/project/tensorflow-zh/api_docs/python/state_ops.html