提交 96f26445 编写于 作者: 究其根本's avatar 究其根本 提交者: Aston Zhang

small typo (#429)

* small typo

* small typo

* code and description is not same

* statements polishing

* statements polishing

* minor typo

* statements polishing

* minor typo

* minor typo

* minor typo

* minor typo

* minor typo

* minor typo

* minor typo

* minor typo

* minor typo

* minor typo

* minor typo

* minor  typo

* minor typo

* minor typo

* minor typo

* minor typos

* GAN typos

* optimization typos

* optimization typos

* sync the latest master branch

* small typo
上级 59ba17b7
......@@ -5,7 +5,7 @@
![深度循环神经网络的架构。](../img/deep-rnn.svg)
具体来说,在时间步$t$里,设小批量输入$\boldsymbol{X}_t \in \mathbb{R}^{n \times d}$(样本数为$n$,输入个数为$d$),第$\ell$隐藏层($\ell=1,\ldots,T$)的隐藏状态为$\boldsymbol{H}_t^{(\ell)} \in \mathbb{R}^{n \times h}$(隐藏单元个数为$h$),输出层变量为$\boldsymbol{O}_t \in \mathbb{R}^{n \times q}$(输出个数为$q$),且隐藏层的激活函数为$\phi$。第1隐藏层的隐藏状态和之前的计算一样:
具体来说,在时间步$t$里,设小批量输入$\boldsymbol{X}_t \in \mathbb{R}^{n \times d}$(样本数为$n$,输入个数为$d$),第$\ell$隐藏层($\ell=1,\ldots,L$)的隐藏状态为$\boldsymbol{H}_t^{(\ell)} \in \mathbb{R}^{n \times h}$(隐藏单元个数为$h$),输出层变量为$\boldsymbol{O}_t \in \mathbb{R}^{n \times q}$(输出个数为$q$),且隐藏层的激活函数为$\phi$。第1隐藏层的隐藏状态和之前的计算一样:
$$\boldsymbol{H}_t^{(1)} = \phi(\boldsymbol{X}_t \boldsymbol{W}_{xh}^{(1)} + \boldsymbol{H}_{t-1}^{(1)} \boldsymbol{W}_{hh}^{(1)} + \boldsymbol{b}_h^{(1)}),$$
......
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