循环神经网络的简洁实现

In [1]:
import torch
from torch import nn
from torch.nn import functional as F
from d2l import torch as d2l

batch_size, num_steps = 32, 35
train_iter, vocab = d2l.load_data_time_machine(batch_size, num_steps)

定义模型

In [2]:
num_hiddens = 256
rnn_layer = nn.RNN(len(vocab), num_hiddens)

使用张量来初始化隐藏状态

In [3]:
state = torch.zeros((1, batch_size, num_hiddens))
state.shape
Out[3]:
torch.Size([1, 32, 256])

通过一个隐藏状态和一个输入,我们可以用更新后的隐藏状态计算输出

In [4]:
X = torch.rand(size=(num_steps, batch_size, len(vocab)))
Y, state_new = rnn_layer(X, state)
Y.shape, state_new.shape
Out[4]:
(torch.Size([35, 32, 256]), torch.Size([1, 32, 256]))

我们为一个完整的循环神经网络模型定义一个RNNModel

In [5]:
class RNNModel(nn.Module):
    """循环神经网络模型。"""
    def __init__(self, rnn_layer, vocab_size, **kwargs):
        super(RNNModel, self).__init__(**kwargs)
        self.rnn = rnn_layer
        self.vocab_size = vocab_size
        self.num_hiddens = self.rnn.hidden_size
        if not self.rnn.bidirectional:
            self.num_directions = 1
            self.linear = nn.Linear(self.num_hiddens, self.vocab_size)
        else:
            self.num_directions = 2
            self.linear = nn.Linear(self.num_hiddens * 2, self.vocab_size)

    def forward(self, inputs, state):
        X = F.one_hot(inputs.T.long(), self.vocab_size)
        X = X.to(torch.float32)
        Y, state = self.rnn(X, state)
        output = self.linear(Y.reshape((-1, Y.shape[-1])))
        return output, state

    def begin_state(self, device, batch_size=1):
        if not isinstance(self.rnn, nn.LSTM):
            return torch.zeros((self.num_directions * self.rnn.num_layers,
                                batch_size, self.num_hiddens), device=device)
        else:
            return (torch.zeros((self.num_directions * self.rnn.num_layers,
                                 batch_size, self.num_hiddens),
                                device=device),
                    torch.zeros((self.num_directions * self.rnn.num_layers,
                                 batch_size, self.num_hiddens),
                                device=device))

用一个具有随机权重的模型进行预测

In [6]:
device = d2l.try_gpu()
net = RNNModel(rnn_layer, vocab_size=len(vocab))
net = net.to(device)
d2l.predict_ch8('time traveller', 10, net, vocab, device)
Out[6]:
'time travellernccccccccc'

使用高级API训练模型

In [7]:
num_epochs, lr = 500, 1
d2l.train_ch8(net, train_iter, vocab, lr, num_epochs, device)
perplexity 1.3, 296747.3 tokens/sec on cuda:0
time travellerit s against reason said filbycan a cube that not 
travellerit s against reason said filbycan a cube that does
2021-07-13T17:17:11.435775 image/svg+xml Matplotlib v3.3.4, https://matplotlib.org/