Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

第六章_循环神经网络(RNN) #28

Merged
merged 22 commits into from
Oct 26, 2018
Merged
Changes from 1 commit
Commits
Show all changes
22 commits
Select commit Hold shift + click to select a range
dcb4263
Add files via upload
tectal Oct 26, 2018
aeb7ebc
Add files via upload
tectal Oct 26, 2018
d93a5d5
Add files via upload
tectal Oct 26, 2018
465fd95
Update 第六章_循环神经网络(RNN).md
tectal Oct 26, 2018
7353e15
Update 第六章_循环神经网络(RNN).md
tectal Oct 26, 2018
905f96c
Update 第六章_循环神经网络(RNN).md
tectal Oct 26, 2018
620bff1
Update 第六章_循环神经网络(RNN).md
tectal Oct 26, 2018
b2dcf0e
Update 第六章_循环神经网络(RNN).md
tectal Oct 26, 2018
b20afbf
Update 第六章_循环神经网络(RNN).md
tectal Oct 26, 2018
65aab75
Update 第六章_循环神经网络(RNN).md
tectal Oct 26, 2018
cdf731b
Update 第六章_循环神经网络(RNN).md
tectal Oct 26, 2018
32daa94
Update 第六章_循环神经网络(RNN).md
tectal Oct 26, 2018
f933858
Update 第六章_循环神经网络(RNN).md
tectal Oct 26, 2018
feb48d6
Update 第六章_循环神经网络(RNN).md
tectal Oct 26, 2018
8a0e78e
Update 第六章_循环神经网络(RNN).md
tectal Oct 26, 2018
6fe099e
Update 第六章_循环神经网络(RNN).md
tectal Oct 26, 2018
e5c98f9
Update 第六章_循环神经网络(RNN).md
tectal Oct 26, 2018
59d74f0
Update 第六章_循环神经网络(RNN).md
tectal Oct 26, 2018
814882e
Update 第六章_循环神经网络(RNN).md
tectal Oct 26, 2018
8c2a039
Update 第六章_循环神经网络(RNN).md
tectal Oct 26, 2018
c9729b9
Update 第六章_循环神经网络(RNN).md
tectal Oct 26, 2018
6d5d96c
Update 第六章_循环神经网络(RNN).md
tectal Oct 26, 2018
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
Prev Previous commit
Next Next commit
Update 第六章_循环神经网络(RNN).md
  • Loading branch information
tectal authored Oct 26, 2018
commit 65aab75a0ddd889f5aa94c96786a2e63560a9051
2 changes: 1 addition & 1 deletion MarkDown/第六章_循环神经网络(RNN).md
Original file line number Diff line number Diff line change
Expand Up @@ -105,4 +105,4 @@ CW-RNNs也是一个RNNs的改良版本,是一种使用时钟频率来驱动的
CW-RNNs与SRNs网络结构类似,也包括输入层(Input)、隐藏层(Hidden)、输出层(Output),它们之间也有向前连接,输入层到隐藏层的连接,隐藏层到输出层的连接。但是与SRN不同的是,隐藏层中的神经元会被划分为若干个组,设为$g$,每一组中的神经元个数相同,设为$k$,并为每一个组分配一个时钟周期$T_i\epsilon\{T_1,T_2,...,T_g\}$,每一个组中的所有神经元都是全连接,但是组$j$到组$i$的循环连接则需要满足$T_j$大于$T_i$。如下图所示,将这些组按照时钟周期递增从左到右进行排序,即$T_1<T_2<...<T_g$,那么连接便是从右到左。例如:隐藏层共有256个节点,分为四组,周期分别是[1,2,4,8],那么每个隐藏层组256/4=64个节点,第一组隐藏层与隐藏层的连接矩阵为64$\times$64的矩阵,第二层的矩阵则为64$\times$128矩阵,第三组为64$\times$(3$\times$64)=64$\times$192矩阵,第四组为64$\times$(4$\times$64)=64$\times$256矩阵。这就解释了上一段的后面部分,速度慢的组连到速度快的组,反之则不成立。

CW-RNNs的网络结构如下图所示:
![](https://github.com/tectal/DeepLearning-500-questions/blob/master/img/ch6/figure_6.6.7_1.png)
![](../img/ch6/figure_6.6.7_1.png)