动手学深度学习Task04

Task04

1.机器翻译及相关技术

机器翻译（MT）：将一段文本从一种语言自动翻译为另一种语言，用神经网络解决这个问题通常称为神经机器翻译（NMT）。
主要特征：输出是单词序列而不是单个单词。 输出序列的长度可能与源序列的长度不同。

import os
os.listdir('/home/kesci/input/')

out：[‘fraeng6506’, ‘d2l9528’, ‘d2l6239’]
引入相关包:

import sys
sys.path.append('/home/kesci/input/d2l9528/')
import collections
import d2l
import zipfile
from d2l.data.base import Vocab
import time
import torch
import torch.nn as nn
import torch.nn.functional as F
from torch.utils import data
from torch import optim

对数据进行预处理，首先看一下数据集的内容：

with open('/home/kesci/input/fraeng6506/fra.txt', 'r') as f:
      raw_text = f.read()
print(raw_text[0:1000])

out：
Go. Va ! CC-BY 2.0 (France) Attribution: tatoeba.org #2877272 (CM) & #1158250 (Wittydev)
Hi. Salut ! CC-BY 2.0 (France) Attribution: tatoeba.org #538123 (CM) & #509819 (Aiji)
Hi. Salut. CC-BY 2.0 (France) Attribution: tatoeba.org #538123 (CM) & #4320462 (gillux)
Run! Cours ! CC-BY 2.0 (France) Attribution: tatoeba.org #906328 (papabear) & #906331 (sacredceltic)
Run! Courez ! CC-BY 2.0 (France) Attribution: tatoeba.org #906328 (papabear) & #906332 (sacredceltic)
Who? Qui ? CC-BY 2.0 (France) Attribution: tatoeba.org #2083030 (CK) & #4366796 (gillux)
Wow! Ça alors ! CC-BY 2.0 (France) Attribution: tatoeba.org #52027 (Zifre) & #374631 (zmoo)
Fire! Au feu ! CC-BY 2.0 (France) Attribution: tatoeba.org #1829639 (Spamster) & #4627939 (sacredceltic)
Help! À l’aide ! CC-BY 2.0 (France) Attribution: tatoeba.org #435084 (lukaszpp) & #128430 (sysko)
Jump. Saute. CC-BY 2.0 (France) Attribution: tatoeba.org #631038 (Shishir) & #2416938 (Phoenix)
Stop! Ça suffit ! CC-BY 2.0 (France) Att

可以看到数据一共由2列有效信息和一些无效信息组成，有效信息的第一列为英文，第二列为对应的法语。

下面对数据进行第一步处理，首先是原文本法语词与其后的符号之间有一个空格，用\xa0表示，为了方便处理需要将其替换成\u202f，然后将所有的大写字母转换成小写字母（如果不区分大小写的话，Go和go代表的含义就不同，翻译时就会出现问题），然后再英语单词与英语单词后的标点之间添加一个空格。

字符在计算机里是以编码的形式存在，我们通常所用的空格是 \x20 ，是在标准ASCII可见字符 0x20~0x7e 范围内。 而 \xa0属于 latin1 （ISO/IEC_8859-1）中的扩展字符集字符，代表不间断空白符nbsp(non-breaking space)，超出gbk编码范围，是需要去除的特殊字符。在数据预处理的过程中，我们首先需要对数据进行清洗。

def preprocess_raw(text):
    text = text.replace('\u202f', ' ').replace('\xa0', ' ')
    out = ''
    for i, char in enumerate(text.lower()):
        if char in (',', '!', '.') and i > 0 and text[i-1] != ' ':
            out += ' '
        out += char
    return out

text = preprocess_raw(raw_text)
print(text[0:1000])

out：
go . va ! cc-by 2 .0 (france) attribution: tatoeba .org #2877272 (cm) & #1158250 (wittydev)
hi . salut ! cc-by 2 .0 (france) attribution: tatoeba .org #538123 (cm) & #509819 (aiji)
hi . salut . cc-by 2 .0 (france) attribution: tatoeba .org #538123 (cm) & #4320462 (gillux)
run ! cours ! cc-by 2 .0 (france) attribution: tatoeba .org #906328 (papabear) & #906331 (sacredceltic)
run ! courez ! cc-by 2 .0 (france) attribution: tatoeba .org #906328 (papabear) & #906332 (sacredceltic)
who? qui ? cc-by 2 .0 (france) attribution: tatoeba .org #2083030 (ck) & #4366796 (gillux)
wow ! ça alors ! cc-by 2 .0 (france) attribution: tatoeba .org #52027 (zifre) & #374631 (zmoo)
fire ! au feu ! cc-by 2 .0 (france) attribution: tatoeba .org #1829639 (spamster) & #4627939 (sacredceltic)
help ! à l’aide ! cc-by 2 .0 (france) attribution: tatoeba .org #435084 (lukaszpp) & #128430 (sysko)
jump . saute . cc-by 2 .0 (france) attribution: tatoeba .org #631038 (shishir) & #2416938 (phoenix)
stop ! ça suffit ! cc-b
下面是进行分词，将英语单词和法语单词分别装入source列表和target列表中。具体实现方法是用换行符’\n’区分出行，再用制表符’\t’区分出英文单词和法文单词，最后把它们分开放到两个列表中。

num_examples = 50000
source, target = [], []
for i, line in enumerate(text.split('\n')):
    if i > num_examples:
        break
    parts = line.split('\t')
    if len(parts) >= 2:
        source.append(parts[0].split(' '))
        target.append(parts[1].split(' '))
source[0:3], target[0:3]

建立词典，即由单词组成的列表转换为单词id组成的列表。
vocab类中包含init、len、getitem方法，这部分在前面已经介绍过了。

def build_vocab(tokens):
    tokens = [token for line in tokens for token in line]
    return d2l.data.base.Vocab(tokens, min_freq=3, use_special_tokens=True)

src_vocab = build_vocab(source)
len(src_vocab)

out：3789
pad函数的作用是对较短序列的token进行补齐，对序列较长的token进行切断，使输入的token长度均为max_len，保证了输入序列的长度一致。

def pad(line, max_len, padding_token):
    if len(line) > max_len:
        return line[:max_len]
    return line + [padding_token] * (max_len - len(line))
pad(src_vocab[source[0]], 10, src_vocab.pad)

out:[38, 4, 0, 0, 0, 0, 0, 0, 0, 0]
可以看到source[0]是[‘go’, ‘.’]，再经过列表转换、pad后变成了[38, 4, 0, 0, 0, 0, 0, 0, 0, 0]。
载入数据集：

def load_data_nmt(batch_size, max_len): # This function is saved in d2l.
    src_vocab, tgt_vocab = build_vocab(source), build_vocab(target)
    src_array, src_valid_len = build_array(source, src_vocab, max_len, True)
    tgt_array, tgt_valid_len = build_array(target, tgt_vocab, max_len, False)
    train_data = data.TensorDataset(src_array, src_valid_len, tgt_array, tgt_valid_len)
    train_iter = data.DataLoader(train_data, batch_size, shuffle=True)
    return src_vocab, tgt_vocab, train_iter

Encoder-Decoder：可以应用在对话系统、生成式任务中。
encoder：输入到隐藏状态
decoder：隐藏状态到输出

encoder-decoder模型的一些理解
定义类Encoder、Decoder以及EncoderDecoder。

class Encoder(nn.Module):
    def __init__(self, **kwargs):
        super(Encoder, self).__init__(**kwargs)

    def forward(self, X, *args):
        raise NotImplementedError
class Decoder(nn.Module):
    def __init__(self, **kwargs):
        super(Decoder, self).__init__(**kwargs)

    def init_state(self, enc_outputs, *args):
        raise NotImplementedError

    def forward(self, X, state):
        raise NotImplementedError
class EncoderDecoder(nn.Module):
    def __init__(self, encoder, decoder, **kwargs):
        super(EncoderDecoder, self).__init__(**kwargs)
        self.encoder = encoder
        self.decoder = decoder

    def forward(self, enc_X, dec_X, *args):
        enc_outputs = self.encoder(enc_X, *args)
        dec_state = self.decoder.init_state(enc_outputs, *args)
        return self.decoder(dec_X, dec_state)

Sequence to Sequence模型

具体结构：

encoder：

class Seq2SeqEncoder(d2l.Encoder):
    def __init__(self, vocab_size, embed_size, num_hiddens, num_layers,
                 dropout=0, **kwargs):
        super(Seq2SeqEncoder, self).__init__(**kwargs)
        self.num_hiddens=num_hiddens
        self.num_layers=num_layers
        self.embedding = nn.Embedding(vocab_size, embed_size)
        self.rnn = nn.LSTM(embed_size,num_hiddens, num_layers, dropout=dropout)
   
    def begin_state(self, batch_size, device):
        return [torch.zeros(size=(self.num_layers, batch_size, self.num_hiddens),  device=device),
                torch.zeros(size=(self.num_layers, batch_size, self.num_hiddens),  device=device)]
    def forward(self, X, *args):
        X = self.embedding(X) # X shape: (batch_size, seq_len, embed_size)
        X = X.transpose(0, 1)  # RNN needs first axes to be time
        # state = self.begin_state(X.shape[1], device=X.device)
        out, state = self.rnn(X)
        # The shape of out is (seq_len, batch_size, num_hiddens).
        # state contains the hidden state and the memory cell
        # of the last time step, the shape is (num_layers, batch_size, num_hiddens)
        return out, state

decoder：

class Seq2SeqDecoder(d2l.Decoder):
    def __init__(self, vocab_size, embed_size, num_hiddens, num_layers,
                 dropout=0, **kwargs):
        super(Seq2SeqDecoder, self).__init__(**kwargs)
        self.embedding = nn.Embedding(vocab_size, embed_size)
        self.rnn = nn.LSTM(embed_size,num_hiddens, num_layers, dropout=dropout)
        self.dense = nn.Linear(num_hiddens,vocab_size)

    def init_state(self, enc_outputs, *args):
        return enc_outputs[1]

    def forward(self, X, state):
        X = self.embedding(X).transpose(0, 1)
        out, state = self.rnn(X, state)
        # Make the batch to be the first dimension to simplify loss computation.
        out = self.dense(out).transpose(0, 1)
        return out, state

定义损失函数：

def SequenceMask(X, X_len,value=0):
    maxlen = X.size(1)
    mask = torch.arange(maxlen)[None, :].to(X_len.device) < X_len[:, None]   
    X[~mask]=value
    return X

class MaskedSoftmaxCELoss(nn.CrossEntropyLoss):
    # pred shape: (batch_size, seq_len, vocab_size)
    # label shape: (batch_size, seq_len)
    # valid_length shape: (batch_size, )
    def forward(self, pred, label, valid_length):
        # the sample weights shape should be (batch_size, seq_len)
        weights = torch.ones_like(label)
        weights = SequenceMask(weights, valid_length).float()
        self.reduction='none'
        output=super(MaskedSoftmaxCELoss, self).forward(pred.transpose(1,2), label)
        return (output*weights).mean(dim=1)
loss = MaskedSoftmaxCELoss()

注意在这里因为我们在前面有pad操作，实际上pad部分的损失函数是无效的，因此在计算loss时要把pad部分去掉，X_len就是tensor的实际大小。
训练：

def train_ch7(model, data_iter, lr, num_epochs, device):  # Saved in d2l
    model.to(device)
    optimizer = optim.Adam(model.parameters(), lr=lr)
    loss = MaskedSoftmaxCELoss()
    tic = time.time()
    for epoch in range(1, num_epochs+1):
        l_sum, num_tokens_sum = 0.0, 0.0
        for batch in data_iter:
            optimizer.zero_grad()
            X, X_vlen, Y, Y_vlen = [x.to(device) for x in batch]
            Y_input, Y_label, Y_vlen = Y[:,:-1], Y[:,1:], Y_vlen-1
            
            Y_hat, _ = model(X, Y_input, X_vlen, Y_vlen)
            l = loss(Y_hat, Y_label, Y_vlen).sum()
            l.backward()

            with torch.no_grad():
                d2l.grad_clipping_nn(model, 5, device)
            num_tokens = Y_vlen.sum().item()
            optimizer.step()
            l_sum += l.sum().item()
            num_tokens_sum += num_tokens
        if epoch % 50 == 0:
            print("epoch {0:4d},loss {1:.3f}, time {2:.1f} sec".format( 
                  epoch, (l_sum/num_tokens_sum), time.time()-tic))
            tic = time.time()

embed_size, num_hiddens, num_layers, dropout = 32, 32, 2, 0.0
batch_size, num_examples, max_len = 64, 1e3, 10
lr, num_epochs, ctx = 0.005, 300, d2l.try_gpu()
src_vocab, tgt_vocab, train_iter = d2l.load_data_nmt(
    batch_size, max_len,num_examples)
encoder = Seq2SeqEncoder(
    len(src_vocab), embed_size, num_hiddens, num_layers, dropout)
decoder = Seq2SeqDecoder(
    len(tgt_vocab), embed_size, num_hiddens, num_layers, dropout)
model = d2l.EncoderDecoder(encoder, decoder)
train_ch7(model, train_iter, lr, num_epochs, ctx)

out：
epoch 50,loss 0.093, time 38.2 sec
epoch 100,loss 0.046, time 37.9 sec
epoch 150,loss 0.032, time 36.8 sec
epoch 200,loss 0.027, time 37.5 sec
epoch 250,loss 0.026, time 37.8 sec
epoch 300,loss 0.025, time 37.3 sec
测试

def translate_ch7(model, src_sentence, src_vocab, tgt_vocab, max_len, device):
    src_tokens = src_vocab[src_sentence.lower().split(' ')]
    src_len = len(src_tokens)
    if src_len < max_len:
        src_tokens += [src_vocab.pad] * (max_len - src_len)
    enc_X = torch.tensor(src_tokens, device=device)
    enc_valid_length = torch.tensor([src_len], device=device)
    # use expand_dim to add the batch_size dimension.
    enc_outputs = model.encoder(enc_X.unsqueeze(dim=0), enc_valid_length)
    dec_state = model.decoder.init_state(enc_outputs, enc_valid_length)
    dec_X = torch.tensor([tgt_vocab.bos], device=device).unsqueeze(dim=0)
    predict_tokens = []
    for _ in range(max_len):
        Y, dec_state = model.decoder(dec_X, dec_state)
        # The token with highest score is used as the next time step input.
        dec_X = Y.argmax(dim=2)
        py = dec_X.squeeze(dim=0).int().item()
        if py == tgt_vocab.eos:
            break
        predict_tokens.append(py)
    return ' '.join(tgt_vocab.to_tokens(predict_tokens))

for sentence in ['Go .', 'Wow !', "I'm OK .", 'I won !']:
    print(sentence + ' => ' + translate_ch7(
        model, sentence, src_vocab, tgt_vocab, max_len, ctx))

out：
Go . => va !
Wow ! => !
I’m OK . => ça va .
I won ! => j’ai gagné !

Beam Search
集束搜索Beam Search
维特比算法：选择整体分数最高的句子（搜索空间太大，如果字典的长度非常大的话，那么查找的状态将会非常多，影响效率）。
贪心搜索(greedy search)：直接选择每个输出的最大概率，直到出现终结符或最大句子长度。
集束搜索(beam search)：是维特比算法的贪心形式。在维特比所有中由于利用动态规划导致当字典较大时效率低，而集束搜索使用beam size参数来限制在每一步保留下来的可能性词的数量。集束搜索是在测试阶段为了获得更好准确性而采取的一种策略，在训练阶段无需使用。
简单greedy search：

维特比算法

2.注意力机制与Seq2seq模型

注意力机制
注意力机制的基本思想和实现原理
注意力机制是一种利用有限的注意力资源从大量信息中快速筛选出高价值信息的手段。
在“编码器—解码器（seq2seq）”⼀节⾥，解码器在各个时间步依赖相同的背景变量（context vector）来获取输⼊序列信息。当编码器为RNN时，背景变量来⾃它最终时间步的隐藏状态。将源序列输入信息以循环单位状态编码，然后将其传递给解码器以生成目标序列。然而这种结构存在着问题，尤其是RNN机制实际中存在长程梯度消失的问题，对于较长的句子，很难寄希望于将输入的序列转化为定长的向量而保存所有的有效信息，所以随着所需翻译句子的长度的增加，这种结构的效果会显著下降。
与此同时，解码的目标词语可能只与原输入的部分词语有关，而并不是与所有的输入有关。例如，当把“Hello world”翻译成“Bonjour le monde”时，“Hello”映射成“Bonjour”，“world”映射成“monde”,“monde”和“Hello”的联系不是很大。在seq2seq模型中，解码器只能隐式地从编码器的最终状态中选择相应的信息。然而，注意力机制可以将这种选择过程显式地建模。

输入包括询问(Query)和Memory两部分，输出output，其中Memory又包括Values和Keys两部分。
注意力机制框架
Attention 是一种通用的带权池化方法，输入由两部分构成：询问（query）和键值对（key-value pairs）。对于一个query来说，attention layer 会与每一个key计算注意力分数并进行权重的归一化，输出的向量则是value的加权求和，而每个key计算的权重与value一一对应。

import math
import torch 
import torch.nn as nn
import os
def file_name_walk(file_dir):
    for root, dirs, files in os.walk(file_dir):
#         print("root", root)  # 当前目录路径
         print("dirs", dirs)  # 当前路径下所有子目录
         print("files", files)  # 当前路径下所有非目录子文件

file_name_walk("/home/kesci/input/fraeng6506")

out：dirs []
files [’_about.txt’, ‘fra.txt’]
Softmax屏蔽
目的是排除padding部分对attention的影响。长度大于等于X_len的部分为1，小于X_len的部分为0，然后在1的部分，也就是对padding的部分，赋值-∞。

def SequenceMask(X, X_len,value=-1e6):
    maxlen = X.size(1)
    #print(X.size(),torch.arange((maxlen),dtype=torch.float)[None, :],'\n',X_len[:, None] )
    mask = torch.arange((maxlen),dtype=torch.float)[None, :] >= X_len[:, None]   
    #print(mask)
    X[mask]=value
    return X

def masked_softmax(X, valid_length):
    # X: 3-D tensor, valid_length: 1-D or 2-D tensor
    softmax = nn.Softmax(dim=-1)
    if valid_length is None:
        return softmax(X)
    else:
        shape = X.shape
        if valid_length.dim() == 1:
            try:
                valid_length = torch.FloatTensor(valid_length.numpy().repeat(shape[1], axis=0))#[2,2,3,3]
            except:
                valid_length = torch.FloatTensor(valid_length.cpu().numpy().repeat(shape[1], axis=0))#[2,2,3,3]
        else:
            valid_length = valid_length.reshape((-1,))
        # fill masked elements with a large negative, whose exp is 0
        X = SequenceMask(X.reshape((-1, shape[-1])), valid_length)
 
        return softmax(X).reshape(shape)

超出2维矩阵的乘法：例如两个三维矩阵相乘，就是以第一维为变量i进行for循环，循环内使后两维张量相乘。X(b,n,m),Y(b,m,k),得到Z(b,n,k)

点积注意力
假设query和keys有相同的维度,通过计算query和key转置的乘积来计算attention score,通常还会除去根号d 减少计算出来的score对维度的依赖性。假设Q(m,d),K(n,d)，则所有mn个score为Q(K转置)/根号d。
下面是具体实现代码：

# Save to the d2l package.
class DotProductAttention(nn.Module): 
    def __init__(self, dropout, **kwargs):
        super(DotProductAttention, self).__init__(**kwargs)
        self.dropout = nn.Dropout(dropout)

    # query: (batch_size, #queries, d)
    # key: (batch_size, #kv_pairs, d)
    # value: (batch_size, #kv_pairs, dim_v)
    # valid_length: either (batch_size, ) or (batch_size, xx)
    def forward(self, query, key, value, valid_length=None):
        d = query.shape[-1]
        # set transpose_b=True to swap the last two dimensions of key
        
        scores = torch.bmm(query, key.transpose(1,2)) / math.sqrt(d)
        attention_weights = self.dropout(masked_softmax(scores, valid_length))
        print("attention_weight\n",attention_weights)
        return torch.bmm(attention_weights, value)

然后进行测试，现在我们创建了两个批，每个批有一个query和10个key-values对。我们通过valid_length指定，对于第一批，我们只关注前2个键-值对，而对于第二批，我们将检查前6个键-值对。因此，尽管这两个批处理具有相同的查询和键值对，但我们获得的输出是不同的。

atten = DotProductAttention(dropout=0)

keys = torch.ones((2,10,2),dtype=torch.float)
values = torch.arange((40), dtype=torch.float).view(1,10,4).repeat(2,1,1)
atten(torch.ones((2,1,2),dtype=torch.float), keys, values, torch.FloatTensor([2, 6]))

attention_weight
tensor([[[0.5000, 0.5000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000,
0.0000, 0.0000]],
out：
[[0.1667, 0.1667, 0.1667, 0.1667, 0.1667, 0.1667, 0.0000, 0.0000,
0.0000, 0.0000]]])
tensor([[[ 2.0000, 3.0000, 4.0000, 5.0000]],

    [[10.0000, 11.0000, 12.0000, 13.0000]]])

引入注意力机制的Seq2seq模型
本节中将注意机制添加到sequence to sequence 模型中，以显式地使用权重聚合states。下图展示encoding 和decoding的模型结构，在时间步为t的时候。此刻attention layer保存着encodering看到的所有信息——即encoding的每一步输出。在decoding阶段，解码器t的时刻的隐藏状态被当作query，encoder的每个时间步的hidden states作为key和value进行attention聚合. Attetion model的输出当作成上下文信息context vector，并与解码器输入Dt拼接起来一起送到解码器。

import sys
sys.path.append('/home/kesci/input/d2len9900')
import d2l

解码器
由于带有注意机制的seq2seq的编码器与之前章节中的Seq2SeqEncoder相同，所以在此处我们只关注解码器。我们添加了一个MLP注意层(MLPAttention)，它的隐藏大小与解码器中的LSTM层相同。然后我们通过从编码器传递三个参数来初始化解码器的状态:

the encoder outputs of all timesteps：encoder输出的各个状态，被用于attetion layer的memory部分，有相同的key和values
the hidden state of the encoder’s final timestep：编码器最后一个时间步的隐藏状态，被用于初始化decoder 的hidden state
the encoder valid length: 编码器的有效长度，借此，注意层不会考虑编码器输出中的填充标记（Paddings）
在解码的每个时间步，我们使用解码器的最后一个RNN层的输出作为注意层的query。然后，将注意力模型的输出与输入嵌入向量连接起来，输入到RNN层。虽然RNN层隐藏状态也包含来自解码器的历史信息，但是attention model的输出显式地选择了enc_valid_len以内的编码器输出，这样attention机制就会尽可能排除其他不相关的信息。

class Seq2SeqAttentionDecoder(d2l.Decoder):
    def __init__(self, vocab_size, embed_size, num_hiddens, num_layers,
                 dropout=0, **kwargs):
        super(Seq2SeqAttentionDecoder, self).__init__(**kwargs)
        self.attention_cell = MLPAttention(num_hiddens,num_hiddens, dropout)
        self.embedding = nn.Embedding(vocab_size, embed_size)
        self.rnn = nn.LSTM(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_len, *args):
        outputs, hidden_state = enc_outputs
#         print("first:",outputs.size(),hidden_state[0].size(),hidden_state[1].size())
        # Transpose outputs to (batch_size, seq_len, hidden_size)
        return (outputs.permute(1,0,-1), hidden_state, enc_valid_len)
        #outputs.swapaxes(0, 1)
        
    def forward(self, X, state):
        enc_outputs, hidden_state, enc_valid_len = state
        #("X.size",X.size())
        X = self.embedding(X).transpose(0,1)
#         print("Xembeding.size2",X.size())
        outputs = []
        for l, x in enumerate(X):
#             print(f"\n{l}-th token")
#             print("x.first.size()",x.size())
            # query shape: (batch_size, 1, hidden_size)
            # select hidden state of the last rnn layer as query
            query = hidden_state[0][-1].unsqueeze(1) # np.expand_dims(hidden_state[0][-1], axis=1)
            # context has same shape as query
#             print("query enc_outputs, enc_outputs:\n",query.size(), enc_outputs.size(), enc_outputs.size())
            context = self.attention_cell(query, enc_outputs, enc_outputs, enc_valid_len)
            # Concatenate on the feature dimension
#             print("context.size:",context.size())
            x = torch.cat((context, x.unsqueeze(1)), dim=-1)
            # Reshape x to (1, batch_size, embed_size+hidden_size)
#             print("rnn",x.size(), len(hidden_state))
            out, hidden_state = self.rnn(x.transpose(0,1), hidden_state)
            outputs.append(out)
        outputs = self.dense(torch.cat(outputs, dim=0))
        return outputs.transpose(0, 1), [enc_outputs, hidden_state,
                                        enc_valid_len]

现在我们可以用注意力模型来测试seq2seq。为了与第9.7节中的模型保持一致，我们对vocab_size、embed_size、num_hiddens和num_layers使用相同的超参数。结果，我们得到了相同的解码器输出形状，但是状态结构改变了。

encoder = d2l.Seq2SeqEncoder(vocab_size=10, embed_size=8,
                            num_hiddens=16, num_layers=2)
# encoder.initialize()
decoder = Seq2SeqAttentionDecoder(vocab_size=10, embed_size=8,
                                  num_hiddens=16, num_layers=2)
X = torch.zeros((4, 7),dtype=torch.long)
print("batch size=4\nseq_length=7\nhidden dim=16\nnum_layers=2\n")
print('encoder output size:', encoder(X)[0].size())
print('encoder hidden size:', encoder(X)[1][0].size())
print('encoder memory size:', encoder(X)[1][1].size())
state = decoder.init_state(encoder(X), None)
out, state = decoder(X, state)
out.shape, len(state), state[0].shape, len(state[1]), state[1][0].shape

训练

import zipfile
import torch
import requests
from io import BytesIO
from torch.utils import data
import sys
import collections

class Vocab(object): # This class is saved in d2l.
  def __init__(self, tokens, min_freq=0, use_special_tokens=False):
    # sort by frequency and token
    counter = collections.Counter(tokens)
    token_freqs = sorted(counter.items(), key=lambda x: x[0])
    token_freqs.sort(key=lambda x: x[1], reverse=True)
    if use_special_tokens:
      # padding, begin of sentence, end of sentence, unknown
      self.pad, self.bos, self.eos, self.unk = (0, 1, 2, 3)
      tokens = ['', '', '', '']
    else:
      self.unk = 0
      tokens = ['']
    tokens += [token for token, freq in token_freqs if freq >= min_freq]
    self.idx_to_token = []
    self.token_to_idx = dict()
    for token in tokens:
      self.idx_to_token.append(token)
      self.token_to_idx[token] = len(self.idx_to_token) - 1
      
  def __len__(self):
    return len(self.idx_to_token)
  
  def __getitem__(self, tokens):
    if not isinstance(tokens, (list, tuple)):
      return self.token_to_idx.get(tokens, self.unk)
    else:
      return [self.__getitem__(token) for token in tokens]
    
  def to_tokens(self, indices):
    if not isinstance(indices, (list, tuple)):
      return self.idx_to_token[indices]
    else:
      return [self.idx_to_token[index] for index in indices]

def load_data_nmt(batch_size, max_len, num_examples=1000):
    """Download an NMT dataset, return its vocabulary and data iterator."""
    # Download and preprocess
    def preprocess_raw(text):
        text = text.replace('\u202f', ' ').replace('\xa0', ' ')
        out = ''
        for i, char in enumerate(text.lower()):
            if char in (',', '!', '.') and text[i-1] != ' ':
                out += ' '
            out += char
        return out 


    with open('/home/kesci/input/fraeng6506/fra.txt', 'r') as f:
      raw_text = f.read()


    text = preprocess_raw(raw_text)

    # Tokenize
    source, target = [], []
    for i, line in enumerate(text.split('\n')):
        if i >= num_examples:
            break
        parts = line.split('\t')
        if len(parts) >= 2:
            source.append(parts[0].split(' '))
            target.append(parts[1].split(' '))

    # Build vocab
    def build_vocab(tokens):
        tokens = [token for line in tokens for token in line]
        return Vocab(tokens, min_freq=3, use_special_tokens=True)
    src_vocab, tgt_vocab = build_vocab(source), build_vocab(target)

    # Convert to index arrays
    def pad(line, max_len, padding_token):
        if len(line) > max_len:
            return line[:max_len]
        return line + [padding_token] * (max_len - len(line))

    def build_array(lines, vocab, max_len, is_source):
        lines = [vocab[line] for line in lines]
        if not is_source:
            lines = [[vocab.bos] + line + [vocab.eos] for line in lines]
        array = torch.tensor([pad(line, max_len, vocab.pad) for line in lines])
        valid_len = (array != vocab.pad).sum(1)
        return array, valid_len

    src_vocab, tgt_vocab = build_vocab(source), build_vocab(target)
    src_array, src_valid_len = build_array(source, src_vocab, max_len, True)
    tgt_array, tgt_valid_len = build_array(target, tgt_vocab, max_len, False)
    train_data = data.TensorDataset(src_array, src_valid_len, tgt_array, tgt_valid_len)
    train_iter = data.DataLoader(train_data, batch_size, shuffle=True)
    return src_vocab, tgt_vocab, train_iter

embed_size, num_hiddens, num_layers, dropout = 32, 32, 2, 0.0
batch_size, num_steps = 64, 10
lr, num_epochs, ctx = 0.005, 500, d2l.try_gpu()

src_vocab, tgt_vocab, train_iter = 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)
model = d2l.EncoderDecoder(encoder, decoder)

训练

d2l.train_s2s_ch9(model, train_iter, lr, num_epochs, ctx)

out：
epoch 50,loss 0.104, time 54.7 sec
epoch 100,loss 0.046, time 54.8 sec
epoch 150,loss 0.031, time 54.7 sec
epoch 200,loss 0.027, time 54.3 sec
epoch 250,loss 0.025, time 54.3 sec
epoch 300,loss 0.024, time 54.4 sec
epoch 350,loss 0.024, time 54.4 sec
epoch 400,loss 0.024, time 54.5 sec
epoch 450,loss 0.023, time 54.4 sec
epoch 500,loss 0.023, time 54.7 sec
预测：

for sentence in ['Go .', 'Good Night !', "I'm OK .", 'I won !']:
    print(sentence + ' => ' + d2l.predict_s2s_ch9(
        model, sentence, src_vocab, tgt_vocab, num_steps, ctx))

Go . => va !
Good Night ! => !
I’m OK . => ça va .
I won ! => j’ai gagné !

3.Transformer

在之前的章节中，我们已经介绍了主流的神经网络架构如卷积神经网络（CNNs）和循环神经网络（RNNs）。
CNNs 易于并行化，却不适合捕捉变长序列内的依赖关系。
RNNs 适合捕捉长距离变长序列的依赖，但是却难以实现并行化处理序列。
下面实现一个Transformer里全新的子结构，并且构建一个神经机器翻译模型用以训练和测试。

复制了上一小节中 masked softmax 实现。

import os
import math
import numpy as np
import torch 
import torch.nn as nn
import torch.nn.functional as F
import sys
sys.path.append('/home/kesci/input/d2len9900')
import d2l
def SequenceMask(X, X_len,value=-1e6):
    maxlen = X.size(1)
    X_len = X_len.to(X.device)
    #print(X.size(),torch.arange((maxlen),dtype=torch.float)[None, :],'\n',X_len[:, None] )
    mask = torch.arange((maxlen), dtype=torch.float, device=X.device)
    mask = mask[None, :] < X_len[:, None]
    #print(mask)
    X[~mask]=value
    return X

def masked_softmax(X, valid_length):
    # X: 3-D tensor, valid_length: 1-D or 2-D tensor
    softmax = nn.Softmax(dim=-1)
    if valid_length is None:
        return softmax(X)
    else:
        shape = X.shape
        if valid_length.dim() == 1:
            try:
                valid_length = torch.FloatTensor(valid_length.numpy().repeat(shape[1], axis=0))#[2,2,3,3]
            except:
                valid_length = torch.FloatTensor(valid_length.cpu().numpy().repeat(shape[1], axis=0))#[2,2,3,3]
        else:
            valid_length = valid_length.reshape((-1,))
        # fill masked elements with a large negative, whose exp is 0
        X = SequenceMask(X.reshape((-1, shape[-1])), valid_length)
 
        return softmax(X).reshape(shape)

# Save to the d2l package.
class DotProductAttention(nn.Module): 
    def __init__(self, dropout, **kwargs):
        super(DotProductAttention, self).__init__(**kwargs)
        self.dropout = nn.Dropout(dropout)

    # query: (batch_size, #queries, d)
    # key: (batch_size, #kv_pairs, d)
    # value: (batch_size, #kv_pairs, dim_v)
    # valid_length: either (batch_size, ) or (batch_size, xx)
    def forward(self, query, key, value, valid_length=None):
        d = query.shape[-1]
        # set transpose_b=True to swap the last two dimensions of key
        scores = torch.bmm(query, key.transpose(1,2)) / math.sqrt(d)
        attention_weights = self.dropout(masked_softmax(scores, valid_length))
        return torch.bmm(attention_weights, value)

多头注意力层
自注意力模型是一个正规的注意力模型，序列的每一个元素对应的key，value，query是完全一致的。自注意力输出了一个与输入长度相同的表征序列，与循环神经网络相比，自注意力对每个元素输出的计算是并行的，所以可以高效的实现这个模块。
多头注意力层包含h个并行的自注意力层，每一个这种层被成为一个head。对每个头来说，在进行注意力计算之前，我们会将query、key和value用三个现行层进行映射，这h个注意力头的输出将会被拼接之后输入最后一个线性层进行整合。