logo
Loading...

神經網路參數說明與數學式表達方式 - Cupoy

先附上 githubhttps://github.com/gn0262487838/100_day_...

神經網路參數說明與數學式表達方式

2020/06/11 08:29 下午
機器學習共學討論版
Jack
觀看數:6
回答數:2
收藏數:1

先附上 github

https://github.com/gn0262487838/100_day_marathon_for_ML/blob/master/D75/Day75-Back_Propagation_Ans.ipynb


其中一段如下:

   # in what direction is the target value?
   
# were we really sure? if so, don't change too much.
   l2_delta
= l2_error*nonlin(l2,deriv=True)

   
# how much did each l1 value contribute to the l2 error (according to the weights)?
   l1_error
= l2_delta.dot(syn1.T)


無法理解這兩個code的意義,能否用數學角度去解釋?