Deep learning II - I Practical aspects of deep learning - Normalizing inputs 輸入歸一化

Normalizing inputs 輸入歸一化 1. 輸入歸一化能夠加快訓練速度 2. 首先，將輸入減去均值 μ=1m∑i=1mx(i) μ = 1 m ∑ i = 1 m x ( i ) x:=x−μ x := x − μ 3. 其次，方差歸一化 σ2=1m∑i=1mx(i)∗∗2(∗∗meanselement−square) σ 2 = 1 m ∑ i = 1 m x ( i ) ∗ ∗ 2

>>阅读原文<<