TextLSTM

nn.LSTM

PyTorch LSTM API 文档

输入数据格式：

• input: [seq_len, batch, input_size]
• : [num_layers * num_directions, batch, hidden_size]
• : [num_layers * num_directions, batch, hidden_size]

输出数据格式：

• output [seq_len, batch, hidden_size * num_directions]
• : [num_layers * num_directions, batch, hidden_size]
• : [num_layers * num_directions, batch, hidden_size]

接下来看个具体的例子

import torch
import torch.nn as nn

lstm = nn.LSTM(input_size=100, hidden_size=20, num_layers=4)
x = torch.randn(10, 3, 100) # 一个句子10个单词，送进去3条句子，每个单词用一个100维的vector表示
out, (h, c) = lstm(x)
print(out.shape, h.shape, c.shape)
# torch.Size([10, 3, 20]) torch.Size([4, 3, 20]) torch.Size([4, 3, 20])

nn.LSTMCell

和 RNNCell 类似，输入 input_size 的 shape 是[batch, input_size]，输出 和 的 shape 是 [batch, hidden_size]

看个一层的 LSTM 的例子

import torch
import torch.nn as nn

cell = nn.LSTMCell(input_size=100, hidden_size=20) # one layer LSTM
h = torch.zeros(3, 20)
c = torch.zeros(3, 20)
x = torch.randn(10, 3, 100)
for xt in x:
    h, c = cell(xt, [h, c])
print(h.shape, c.shape) # torch.Size([3, 20]) torch.Size([3, 20])

两层的LSTM例子

import torch
import torch.nn as nn

cell1 = nn.LSTMCell(input_size=100, hidden_size=30)
cell2 = nn.LSTMCell(input_size=30, hidden_size=20)
h1 = torch.zeros(3, 30)
c1 = torch.zeros(3, 30)
h2 = torch.zeros(3, 20)
c2 = torch.zeros(3, 20)
x = torch.randn(10, 3, 100)
for xt in x:
    h1, c1 = cell1(xt, [h1, c1])
    h2, c2 = cell2(h1, [h2, c2])
print(h2.shape, c2.shape) # torch.Size([3, 20]) torch.Size([3, 20])