32.Bahdanau 注意力

python 复制代码

import torch
from torch import nn
from d2l import torch as d2l
########################################################################################################################################
#@save
class AttentionDecoder(d2l.Decoder):
    """带有注意力机制解码器的基本接口"""
    def __init__(self, **kwargs):
        super(AttentionDecoder, self).__init__(**kwargs)

    @property
    def attention_weights(self):
        raise NotImplementedError
########################################################################################################################################
class Seq2SeqAttentionDecoder(AttentionDecoder):
    def __init__(self,vocab_size,embed_size,num_hiddens,num_layers,
                 dropout=0,**kwargs):
        super(Seq2SeqAttentionDecoder,self).__init__(**kwargs)
        self.attention=d2l.AdditiveAttention(num_hiddens,num_hiddens,num_hiddens,dropout)
        self.embedding=nn.Embedding(vocab_size,embed_size)
        self.rnn=nn.GRU(embed_size+num_hiddens,num_hiddens,num_layers,dropout=dropout)
        self.dense=nn.Linear(num_hiddens,vocab_size)
    def init_state(self, enc_outputs, enc_valid_lens, *args):
        #enc_outputs:[b,t,h]
        #hidden_state:[num_layers,b,h]
        outputs, hidden_state = enc_outputs
        return (outputs.permute(1, 0, 2), hidden_state, enc_valid_lens)
    def forward(self, X, state):
        enc_outputs,hidden_state,enc_valid_lens=state
        #[B,T,E]->[T,B,E]
        X=self.embedding(X).permute(1,0,2)
        outputs,self._attention_weights=[],[]
        for x in X:
            #[B,1,H]
            query=torch.unsqueeze(hidden_state[-1],dim=1)
            #[B,1,H]
            context=self.attention(query,enc_outputs,enc_outputs,enc_valid_lens)
            #[B,1,H+H]
            x=torch.cat((context,torch.unsqueeze(x,dim=1)),dim=-1)
            out,hidden_state=self.rnn(x.permute(1,0,2),hidden_state)
            outputs.append(out)
            self._attention_weights.append(self.attention.attention_weights)
        #(T,B,V)
        outputs=self.dense(torch.cat(outputs,dim=0))
        return outputs.permute(1,0,2),[enc_outputs,hidden_state,enc_valid_lens]
    @property
    def attention_weights(self):
        return self._attention_weights
########################################################################################################################################
encoder = d2l.Seq2SeqEncoder(vocab_size=10, embed_size=8, num_hiddens=16,num_layers=2)
encoder.eval()
decoder = Seq2SeqAttentionDecoder(vocab_size=10, embed_size=8, num_hiddens=16,num_layers=2,dropout=0.1)
decoder.eval()
X = torch.zeros((4, 7), dtype=torch.long)  # (batch_size,num_steps)
state = decoder.init_state(encoder(X), None)
output, state = decoder(X, state)
#参数测试:
embed_size, num_hiddens, num_layers, dropout = 32, 32, 2, 0.1
batch_size, num_steps = 64, 10
lr, num_epochs, device = 0.005, 250, d2l.try_gpu()

train_iter, src_vocab, tgt_vocab = d2l.load_data_nmt(batch_size, num_steps)
encoder = d2l.Seq2SeqEncoder(
    len(src_vocab), embed_size, num_hiddens, num_layers, dropout)
decoder = Seq2SeqAttentionDecoder(
    len(tgt_vocab), embed_size, num_hiddens, num_layers, dropout)
net = d2l.EncoderDecoder(encoder, decoder)
d2l.train_seq2seq(net, train_iter, lr, num_epochs, tgt_vocab, device)
########################################################################################################################################

python 复制代码

engs = ['go .', "i lost .", 'he\'s calm .', 'i\'m home .']
fras = ['va !', 'j\'ai perdu .', 'il est calme .', 'je suis chez moi .']
for eng, fra in zip(engs, fras):
    translation, dec_attention_weight_seq = d2l.predict_seq2seq(
        net, eng, src_vocab, tgt_vocab, num_steps, device, True)
    print(f'{eng} => {translation}, ',
          f'bleu {d2l.bleu(translation, fra, k=2):.3f}')
    
attention_weights = torch.cat([step[0][0][0] for step in dec_attention_weight_seq], 0).reshape((
    1, 1, -1, num_steps))
d2l.show_heatmaps(
    attention_weights[:, :, :, :len(engs[-1].split()) + 1].cpu(),
    xlabel='Key positions', ylabel='Query positions')