learning rate decay

在训练神经网络的时候，通常在训练刚开始的时候使用较大的learning rate， 随着训练的进行，我们会慢慢的减小learning rate。对于这种常用的训练策略，tensorflow 也提供了相应的API让我们可以更简单的将这个方法应用到我们训练网络的过程中。

接口
tf.train.exponential_decay(learning_rate, global_step, decay_steps, decay_rate, staircase=False, name=None)
参数:
learning_rate : 初始的learning rate
global_step : 全局的step，与 decay_step 和 decay_rate一起决定了 learning rate的变化。
staircase : 如果为 True global_step/decay_step 向下取整

更新公式：

decayed_learning_rate = learning_rate *
                        decay_rate ^ (global_step / decay_steps)

这个代码可以看一下 learning_rate 的变化趋势：

import tensorflow as tf

global_step = tf.Variable(0, trainable=False)

initial_learning_rate = 0.1 #初始学习率

learning_rate = tf.train.exponential_decay(initial_learning_rate,
                                           global_step=global_step,
                                           decay_steps=10,decay_rate=0.9)
opt = tf.train.GradientDescentOptimizer(learning_rate)

add_global = global_step.assign_add(1)
with tf.Session() as sess:
    tf.global_variables_initializer().run()
    print(sess.run(learning_rate))
    for i in range(10):
        _, rate = sess.run([add_global, learning_rate])
        print(rate)

用法：

import tensorflow as tf

global_step = tf.Variable(0, trainable=False)

initial_learning_rate = 0.1 #初始学习率

learning_rate = tf.train.exponential_decay(initial_learning_rate,
                                           global_step=global_step,
                                           decay_steps=10,decay_rate=0.9)
opt = tf.train.GradientDescentOptimizer(learning_rate)

add_global = global_step.assign_add(1)
with tf.control_denpendices([add_global]):
    train_op = opt.minimise(loss)

with tf.Session() as sess:
    tf.global_variables_initializer().run()
    print(sess.run(learning_rate))
    for i in range(10):
        _= sess.run(train_op)
        print(rate)

参考资料

https://www.tensorflow.org/api_docs/python/tf/train/exponential_decay