12
15

Delete article

Deleted articles cannot be recovered.

Draft of this article would be also deleted.

Are you sure you want to delete this article?

More than 5 years have passed since last update.

Learning Rate調整の基本

Last updated at Posted at 2018-07-15

Learning Rate

Learning Rate(学習率)はハイパーパラメータの中で最も重要なものの一つ。

一般的な値

  • 0.1
  • 0.01
  • 0.001
  • 0.0001
  • 0.00001
  • 0.000001

初期値は0.01にしてみるのが基本。

調整方法

validation errorがより少なくなるように設定する。
validation errorの減少するスピードが遅ければ(①)learning rateを増やし、validation errorが増加してしまっているなら(②)learning rateを減らすなど。
learning rate image

より高度な設定

効率的に学習をすすめるにはLearning Rateをepoch数に応じて減少させる。
[参考]横軸はepoch数
learning rate image2

学習アルゴリズムを使う。

  1. TensorFlow - AdamOptimizer
  2. TensorFlow - AdagradOptimizer
  3. Tensorflow - GradientDescentOptimizer
  4. TensorFlow - MomentumOptimizer
12
15
0

Register as a new user and use Qiita more conveniently

  1. You get articles that match your needs
  2. You can efficiently read back useful information
  3. You can use dark theme
What you can do with signing up
12
15

Delete article

Deleted articles cannot be recovered.

Draft of this article would be also deleted.

Are you sure you want to delete this article?