未验证 提交 67af6b02 编写于 作者: G Garen Wang 提交者: GitHub

Details Fixed in Step 3 (#612)

上级 f67cb3e8
......@@ -6,5 +6,5 @@ from enum import Enum
class NetType(Enum):
Fitting = 1,
BinaryClassifier = 2,
MultipleClassifier = 3
MultipleClassifier = 3,
BinaryTanh = 4
......@@ -226,4 +226,4 @@ A= [[0.65791301]
### 代码位置
ch06, Level1
ch06, Level2
......@@ -160,4 +160,4 @@ b12= -0.49506282174794675
### 代码位置
ch06, Level2
ch06, Level3
......@@ -95,7 +95,7 @@ $$
反向传播误差矩阵为:
$$a-y=[0.879,0.881,0.002]$$
$$a-y=[0.879,-0.881,0.002]$$
本来是第二类,误判为第一类,所以前两个元素的值很大,反向传播的力度就大。
......
Markdown is supported
0% .
You are about to add 0 people to the discussion. Proceed with caution.
先完成此消息的编辑!
想要评论请 注册