tensorflow exponential_decay使用

发表于 | 阅读次数: 
1
2
3
4
5
6
7
8
exponential_decay(
    learning_rate,
    global_step,
    decay_steps,
    decay_rate,
    staircase=False,
    name=None
)

一般用法

1
2
3
4
5
6
learning_rate = tf.train.exponential_decay(
    LEARNING_RATE_BASE,
    global_step,
    mnist.train.num_examples / BATCH_SIZE,
    LEARNING_RATE_DECAY
)

因为每次train会使用BATCH_SIZE多的数据

请作者喝一杯咖啡☕️