PyTorch 深度学习之循环神经网络(基础篇)Basic RNN(十一)

0.Revision: DNN dense 重义层 全连接

RNN处理带有序列的数据

1. What is RNNs?

linear layer

1.1 What is RNN?

tanh (-1, +1)

1.2 RNN Cell in PyTorch

1.3 How to use RNNCell

*先把维度搞清楚 多了一个序列的维度

2. How to use RNN

2.1 How to use RNN - numLayers

2.2 How to use RNN

如果有batch_first:

3. Example 12-1:Using RNNCell

seq->seq

独热向量

Code-Prepare Data

Design Model

Loss and Optimizer

Training Cycle

Result

4. Example 12-2:Using RNN

Change Model

Result

5. Associate a vetcor with a word/character

嵌入层

5.1 One hot vs embedding

数据降维

5.2 Embedding in PyTorch

6. Example 12-3:Using embedding and linear layer

7. Exercise 12 - 1 Using LSTM

8. Exercise 12 - 1 Using GRU

知识点:

1.序列Data

2.循环共享