pytorch做seq2seq注意力模型的翻译

以下是对pytorch 1.0版本 的seq2seq+注意力模型做法语--英语翻译的理解（这个代码在pytorch0.4上也可以正常跑）：

 # -*- coding: utf-8 -*-
 """
 Translation with a Sequence to Sequence Network and Attention
 *************************************************************
 **Author**: `Sean Robertson <https://github.com/spro/practical-pytorch>`_

 In this project we will be teaching a neural network to translate from
 French to English.

 ::

     [KEY: > input, = target, < output]

     > il est en train de peindre un tableau .
     = he is painting a picture .
     < he is painting a picture .

     > pourquoi ne pas essayer ce vin delicieux ?
     = why not try that delicious wine ?
     < why not try that delicious wine ?

     > elle n est pas poete mais romanciere .
     = she is not a poet but a novelist .
     < she not not a poet but a novelist .

     > vous etes trop maigre .
     = you re too skinny .
     < you re all alone .

 ... to varying degrees of success.

 This is made possible by the simple but powerful idea of the `sequence
 to sequence network <http://arxiv.org/abs/1409.3215>`__, in which two
 recurrent neural networks work together to transform one sequence to
 another. An encoder network condenses an input sequence into a vector,
 and a decoder network unfolds that vector into a new sequence.

 .. figure:: /_static/img/seq-seq-images/seq2seq.png
    :alt:

 To improve upon this model we'll use an `attention
 mechanism <https://arxiv.org/abs/1409.0473>`__, which lets the decoder
 learn to focus over a specific range of the input sequence.

 **Recommended Reading:**

 I assume you have at least installed PyTorch, know Python, and
 understand Tensors:

 -  https://pytorch.org/ For installation instructions
 -  :doc:`/beginner/deep_learning_60min_blitz` to get started with PyTorch in general
 -  :doc:`/beginner/pytorch_with_examples` for a wide and deep overview
 -  :doc:`/beginner/former_torchies_tutorial` if you are former Lua Torch user

 It would also be useful to know about Sequence to Sequence networks and
 how they work:

 -  `Learning Phrase Representations using RNN Encoder-Decoder for
    Statistical Machine Translation <http://arxiv.org/abs/1406.1078>`__
 -  `Sequence to Sequence Learning with Neural
    Networks <http://arxiv.org/abs/1409.3215>`__
 -  `Neural Machine Translation by Jointly Learning to Align and
    Translate <https://arxiv.org/abs/1409.0473>`__
 -  `A Neural Conversational Model <http://arxiv.org/abs/1506.05869>`__

 You will also find the previous tutorials on
 :doc:`/intermediate/char_rnn_classification_tutorial`
 and :doc:`/intermediate/char_rnn_generation_tutorial`
 helpful as those concepts are very similar to the Encoder and Decoder
 models, respectively.

 And for more, read the papers that introduced these topics:

 -  `Learning Phrase Representations using RNN Encoder-Decoder for
    Statistical Machine Translation <http://arxiv.org/abs/1406.1078>`__
 -  `Sequence to Sequence Learning with Neural
    Networks <http://arxiv.org/abs/1409.3215>`__
 -  `Neural Machine Translation by Jointly Learning to Align and
    Translate <https://arxiv.org/abs/1409.0473>`__
 -  `A Neural Conversational Model <http://arxiv.org/abs/1506.05869>`__

 **Requirements**
 """
 from __future__ import unicode_literals, print_function, division
 from io import open
 import unicodedata
 import string
 import re
 import random

 import torch
 import torch.nn as nn
 from torch import optim
 import torch.nn.functional as F

 device = torch.device("cuda" if torch.cuda.is_available() else "cpu")

 ######################################################################
 # Loading data files
 # ==================
 #
 # The data for this project is a set of many thousands of English to
 # French translation pairs.
 #
 # `This question on Open Data Stack
 # Exchange <http://opendata.stackexchange.com/questions/3888/dataset-of-sentences-translated-into-many-languages>`__
 # pointed me to the open translation site http://tatoeba.org/ which has
 # downloads available at http://tatoeba.org/eng/downloads - and better
 # yet, someone did the extra work of splitting language pairs into
 # individual text files here: http://www.manythings.org/anki/
 #
 # The English to French pairs are too big to include in the repo, so
 # download to ``data/eng-fra.txt`` before continuing. The file is a tab
 # separated list of translation pairs:
 #
 # ::
 #
 #     I am cold.    J'ai froid.
 #
 # .. Note::
 #    Download the data from
 #    `here <https://download.pytorch.org/tutorial/data.zip>`_
 #    and extract it to the current directory.

 ######################################################################
 # Similar to the character encoding used in the character-level RNN
 # tutorials, we will be representing each word in a language as a one-hot
 # vector, or giant vector of zeros except for a single one (at the index
 # of the word). Compared to the dozens of characters that might exist in a
 # language, there are many many more words, so the encoding vector is much
 # larger. We will however cheat a bit and trim the data to only use a few
 # thousand words per language.
 #
 # .. figure:: /_static/img/seq-seq-images/word-encoding.png
 #    :alt:
 #
 #

 ######################################################################
 # We'll need a unique index per word to use as the inputs and targets of
 # the networks later. To keep track of all this we will use a helper class
 # called ``Lang`` which has word → index (``word2index``) and index → word
 # (``index2word``) dictionaries, as well as a count of each word
 # ``word2count`` to use to later replace rare words.
 #

 SOS_token = 0
 EOS_token = 1

 # 每个单词需要对应唯一的索引作为稍后的网络输入和目标.为了追踪这些索引
 # 则使用一个帮助类 Lang ，类中有 词 → 索引 (word2index) 和 索引 → 词
 # (index2word) 的字典, 以及每个词word2count 用来替换稀疏词汇.

 # 此处创建的Lang 对象来表示源/目标语言，它包含三部分：word2index、
 # index2word 和word2count，分别表示单词到id、id 到单词和单词的词频。
 # word2count的作用是用于过滤一些低频词（把它变成unknown）

 class Lang:
     def __init__(self, name):
         self.name = name
         self.word2index = {}
         self.word2count = {}
         self.index2word = {0: "SOS", 1: "EOS"}
         self.n_words = 2  # Count SOS and EOS

     def addSentence(self, sentence):
         for word in sentence.split(' '):
             self.addWord(word)  # 用于添加单词

     def addWord(self, word):
         if word not in self.word2index:  # 是不是新的词
             # 如果不在word2index里，则需要新的定义字典
             self.word2index[word] = self.n_words
             self.word2count[word] = 1
             self.index2word[self.n_words] = word
             self.n_words += 1  # 相当于每次index+1
         else:
             self.word2count[word] += 1  # 计算每次词的个数

 ######################################################################
 # The files are all in Unicode, to simplify we will turn Unicode
 # characters to ASCII, make everything lowercase, and trim most
 # punctuation.
 #

 # Turn a Unicode string to plain ASCII, thanks to
 # http://stackoverflow.com/a/518232/2809427

 # 此处是为了将Unicode字符串转换为纯ASCII
 # 原文件是Unicode编码
 def unicodeToAscii(s):
     return ''.join(
         c for c in unicodedata.normalize('NFD', s)
         if unicodedata.category(c) != 'Mn'
     )

 # Lowercase, trim, and remove non-letter characters

 # 小写,修剪和删除非字母字符
 def normalizeString(s):
     s = unicodeToAscii(s.lower().strip())
     s = re.sub(r"([.!?])", r" \1", s)
     s = re.sub(r"[^a-zA-Z.!?]+", r" ", s)
     return s

 ######################################################################
 # To read the data file we will split the file into lines, and then split
 # lines into pairs. The files are all English → Other Language, so if we
 # want to translate from Other Language → English I added the ``reverse``
 # flag to reverse the pairs.
 #

 # 要读取数据文件,我们将把文件分成行,然后将行成对分开. 这些文件
 # 都是英文→其他语言,所以如果我们想从其他语言翻译→英文,我们添加了
 # 翻转标志 reverse来翻转词语对.
 def readLangs(lang1, lang2, reverse=False):
     print("Reading lines...")

     # Read the file and split into lines
     # 读取文件并按行分开
     lines = open('data/%s-%s.txt' % (lang1, lang2), encoding='utf-8'). \
         read().strip().split('\n')

     # Split every line into pairs and normalize
     # 将每一行分成两列并进行标准化
     pairs = [[normalizeString(s) for s in l.split('\t')] for l in lines]

     # Reverse pairs, make Lang instances
     # 翻转对,Lang实例化
     if reverse:
         pairs = [list(reversed(p)) for p in pairs]
         input_lang = Lang(lang2)
         output_lang = Lang(lang1)
     else:
         input_lang = Lang(lang1)
         output_lang = Lang(lang2)

     return input_lang, output_lang, pairs

 ######################################################################
 # Since there are a *lot* of example sentences and we want to train
 # something quickly, we'll trim the data set to only relatively short and
 # simple sentences. Here the maximum length is 10 words (that includes
 # ending punctuation) and we're filtering to sentences that translate to
 # the form "I am" or "He is" etc. (accounting for apostrophes replaced
 # earlier).
 #

 # 由于例句较多,为了方便快速训练,则会将数据集裁剪为相对简短的句子.
 # 这里的单词的最大长度是10词(包括结束标点符号),
 # 保留”I am” 和”He is” 开头的数据

 MAX_LENGTH = 10

 eng_prefixes = (
     "i am ", "i m ",
     "he is", "he s ",
     "she is", "she s",
     "you are", "you re ",
     "we are", "we re ",
     "they are", "they re "
 )

 def filterPair(p):
     return len(p[0].split(' ')) < MAX_LENGTH and \
            len(p[1].split(' ')) < MAX_LENGTH and \
            p[1].startswith(eng_prefixes)
     # 是否满足长度

 def filterPairs(pairs):
     return [pair for pair in pairs if filterPair(pair)]

 ######################################################################
 # The full process for preparing the data is:
 #
 # -  Read text file and split into lines, split lines into pairs
 # -  Normalize text, filter by length and content
 # -  Make word lists from sentences in pairs
 #

 def prepareData(lang1, lang2, reverse=False):
     input_lang, output_lang, pairs = readLangs(lang1, lang2, reverse)
     # 读入数据lang1,lang2,并翻转
     print("Read %s sentence pairs" % len(pairs))
     # 一共读入了多少对
     pairs = filterPairs(pairs)
     # 符合条件的配对有多少对
     print("Trimmed to %s sentence pairs" % len(pairs))
     print("Counting words...")
     for pair in pairs:
         input_lang.addSentence(pair[0])
         output_lang.addSentence(pair[1])
     print("Counted words:")
     print(input_lang.name, input_lang.n_words)
     print(output_lang.name, output_lang.n_words)
     return input_lang, output_lang, pairs

 # 对数据进行预处理
 input_lang, output_lang, pairs = prepareData('eng', 'fra', True)
 print(random.choice(pairs))  # 随机展示一对

 ######################################################################
 # The Seq2Seq Model
 # =================
 #
 # A Recurrent Neural Network, or RNN, is a network that operates on a
 # sequence and uses its own output as input for subsequent steps.
 #
 # A `Sequence to Sequence network <http://arxiv.org/abs/1409.3215>`__, or
 # seq2seq network, or `Encoder Decoder
 # network <https://arxiv.org/pdf/1406.1078v3.pdf>`__, is a model
 # consisting of two RNNs called the encoder and decoder. The encoder reads
 # an input sequence and outputs a single vector, and the decoder reads
 # that vector to produce an output sequence.
 #
 # .. figure:: /_static/img/seq-seq-images/seq2seq.png
 #    :alt:
 #
 # Unlike sequence prediction with a single RNN, where every input
 # corresponds to an output, the seq2seq model frees us from sequence
 # length and order, which makes it ideal for translation between two
 # languages.
 #
 # Consider the sentence "Je ne suis pas le chat noir" → "I am not the
 # black cat". Most of the words in the input sentence have a direct
 # translation in the output sentence, but are in slightly different
 # orders, e.g. "chat noir" and "black cat". Because of the "ne/pas"
 # construction there is also one more word in the input sentence. It would
 # be difficult to produce a correct translation directly from the sequence
 # of input words.
 #
 # With a seq2seq model the encoder creates a single vector which, in the
 # ideal case, encodes the "meaning" of the input sequence into a single
 # vector — a single point in some N dimensional space of sentences.
 #

 ######################################################################
 # The Encoder
 # -----------
 #
 # The encoder of a seq2seq network is a RNN that outputs some value for
 # every word from the input sentence. For every input word the encoder
 # outputs a vector and a hidden state, and uses the hidden state for the
 # next input word.
 #
 # .. figure:: /_static/img/seq-seq-images/encoder-network.png
 #    :alt:
 #
 #

 class EncoderRNN(nn.Module):
     def __init__(self, input_size, hidden_size):
         super(EncoderRNN, self).__init__()
         self.hidden_size = hidden_size
         # 定义隐藏层
         self.embedding = nn.Embedding(input_size, hidden_size)
         # word embedding的定义可以这么理解，例如nn.Embedding(2, 4)
         # 2表示有2个词，4表示4维度，其实也就是一个2x4的矩阵，
         # 如果有100个词，每个词10维，就可以写为nn.Embedding(100, 10)
         # 注意这里的词向量的建立只是初始的词向量，并没有经过任何修改优化
         # 需要建立神经网络通过learning的办法修改word embedding里面的参数
         # 使得word embedding每一个词向量能够表示每一个不同的词。
         self.gru = nn.GRU(hidden_size, hidden_size)  # 用到了上面提到的GRU模型

     def forward(self, input, hidden):
         embedded = self.embedding(input).view(1, 1, -1)  # -1是指自适应，view相当于reshape函数
         output = embedded
         output, hidden = self.gru(output, hidden)
         return output, hidden

     def initHidden(self):  # 初始化
         return torch.zeros(1, 1, self.hidden_size, device=device)

 ######################################################################
 # The Decoder
 # -----------
 #
 # The decoder is another RNN that takes the encoder output vector(s) and
 # outputs a sequence of words to create the translation.
 #

 ######################################################################
 # Simple Decoder
 # ^^^^^^^^^^^^^^
 #
 # In the simplest seq2seq decoder we use only last output of the encoder.
 # This last output is sometimes called the *context vector* as it encodes
 # context from the entire sequence. This context vector is used as the
 # initial hidden state of the decoder.
 #
 # At every step of decoding, the decoder is given an input token and
 # hidden state. The initial input token is the start-of-string ``<SOS>``
 # token, and the first hidden state is the context vector (the encoder's
 # last hidden state).
 #
 # .. figure:: /_static/img/seq-seq-images/decoder-network.png
 #    :alt:
 #
 #

 class DecoderRNN(nn.Module):
     # DecoderRNN与encoderRNN结构类似，结合图片即可搞清逻辑
     def __init__(self, hidden_size, output_size):
         super(DecoderRNN, self).__init__()
         self.hidden_size = hidden_size

         self.embedding = nn.Embedding(output_size, hidden_size)
         self.gru = nn.GRU(hidden_size, hidden_size)
         self.out = nn.Linear(hidden_size, output_size)
         self.softmax = nn.LogSoftmax(dim=1)

     def forward(self, input, hidden):
         output = self.embedding(input).view(1, 1, -1)  # -1是指自适应，view相当于reshape函数
         output = F.relu(output)
         output, hidden = self.gru(output, hidden)  # 此处使用gru神经网络
         # 对上述结果使用softmax,就是图片中左边倒数第二个
         output = self.softmax(self.out(output[0]))
         return output, hidden

     def initHidden(self):
         return torch.zeros(1, 1, self.hidden_size, device=device)

 ######################################################################
 # I encourage you to train and observe the results of this model, but to
 # save space we'll be going straight for the gold and introducing the
 # Attention Mechanism.
 #

 ######################################################################
 # Attention Decoder
 # ^^^^^^^^^^^^^^^^^
 #
 # If only the context vector is passed betweeen the encoder and decoder,
 # that single vector carries the burden of encoding the entire sentence.
 #
 # Attention allows the decoder network to "focus" on a different part of
 # the encoder's outputs for every step of the decoder's own outputs. First
 # we calculate a set of *attention weights*. These will be multiplied by
 # the encoder output vectors to create a weighted combination. The result
 # (called ``attn_applied`` in the code) should contain information about
 # that specific part of the input sequence, and thus help the decoder
 # choose the right output words.
 #
 # .. figure:: https://i.imgur.com/1152PYf.png
 #    :alt:
 #
 # Calculating the attention weights is done with another feed-forward
 # layer ``attn``, using the decoder's input and hidden state as inputs.
 # Because there are sentences of all sizes in the training data, to
 # actually create and train this layer we have to choose a maximum
 # sentence length (input length, for encoder outputs) that it can apply
 # to. Sentences of the maximum length will use all the attention weights,
 # while shorter sentences will only use the first few.
 #
 # .. figure:: /_static/img/seq-seq-images/attention-decoder-network.png
 #    :alt:
 #
 #

 class AttnDecoderRNN(nn.Module):
     def __init__(self, hidden_size, output_size, dropout_p=0.1, max_length=MAX_LENGTH):
         super(AttnDecoderRNN, self).__init__()
         self.hidden_size = hidden_size
         self.output_size = output_size
         self.dropout_p = dropout_p
         self.max_length = max_length

         self.embedding = nn.Embedding(self.output_size, self.hidden_size)
         self.attn = nn.Linear(self.hidden_size * 2, self.max_length)
         self.attn_combine = nn.Linear(self.hidden_size * 2, self.hidden_size)
         self.dropout = nn.Dropout(self.dropout_p)
         self.gru = nn.GRU(self.hidden_size, self.hidden_size)
         self.out = nn.Linear(self.hidden_size, self.output_size)

     def forward(self, input, hidden, encoder_outputs):
         # 对于输入的input内容进行embedding和dropout操作
         # dropout是指随机丢弃一些神经元
         embedded = self.embedding(input).view(1, 1, -1)
         embedded = self.dropout(embedded)

         # 此处相当于学出来了attention的权重
         # 需要注意的是torch的concatenate函数是torch.cat，是在已有的维度上拼接，
         # 而stack是建立一个新的维度，然后再在该纬度上进行拼接。
         attn_weights = F.softmax(
             self.attn(torch.cat((embedded[0], hidden[0]), 1)), dim=1)

         # 将attention权重作用在encoder_outputs上
         # 对存储在两个批batch1和batch2内的矩阵进行批矩阵乘操作。
         # batch1和 batch2都为包含相同数量矩阵的3维张量。
         # 如果batch1是形为b×n×m的张量，batch1是形为b×m×p的张量，
         # 则out和mat的形状都是n×p
         attn_applied = torch.bmm(attn_weights.unsqueeze(0),
                                  encoder_outputs.unsqueeze(0))
         # 拼接操作，将embedded和attn_Applied拼接起来
         output = torch.cat((embedded[0], attn_applied[0]), 1)
         # 返回一个新的张量，对输入的制定位置插入维度 1
         output = self.attn_combine(output).unsqueeze(0)

         output = F.relu(output)
         output, hidden = self.gru(output, hidden)

         output = F.log_softmax(self.out(output[0]), dim=1)
         return output, hidden, attn_weights

     def initHidden(self):
         return torch.zeros(1, 1, self.hidden_size, device=device)

 ######################################################################
 # .. note:: There are other forms of attention that work around the length
 #   limitation by using a relative position approach. Read about "local
 #   attention" in `Effective Approaches to Attention-based Neural Machine
 #   Translation <https://arxiv.org/abs/1508.04025>`__.
 #
 # Training
 # ========
 #
 # Preparing Training Data
 # -----------------------
 #
 # To train, for each pair we will need an input tensor (indexes of the
 # words in the input sentence) and target tensor (indexes of the words in
 # the target sentence). While creating these vectors we will append the
 # EOS token to both sequences.
 #

 def indexesFromSentence(lang, sentence):
     return [lang.word2index[word] for word in sentence.split(' ')]

 def tensorFromSentence(lang, sentence):
     # 获得词的索引
     indexes = indexesFromSentence(lang, sentence)
     # 将EOS标记添加到两个序列中
     indexes.append(EOS_token)
     return torch.tensor(indexes, dtype=torch.long, device=device).view(-1, 1)

 def tensorsFromPair(pair):
     # 每一对为需要输入的张量（输入句子中的词的索引）和目标张量
     # （目标语句中的词的索引）
     input_tensor = tensorFromSentence(input_lang, pair[0])
     target_tensor = tensorFromSentence(output_lang, pair[1])
     return (input_tensor, target_tensor)

 ######################################################################
 # Training the Model
 # ------------------
 #
 # To train we run the input sentence through the encoder, and keep track
 # of every output and the latest hidden state. Then the decoder is given
 # the ``<SOS>`` token as its first input, and the last hidden state of the
 # encoder as its first hidden state.
 #
 # "Teacher forcing" is the concept of using the real target outputs as
 # each next input, instead of using the decoder's guess as the next input.
 # Using teacher forcing causes it to converge faster but `when the trained
 # network is exploited, it may exhibit
 # instability <http://minds.jacobs-university.de/sites/default/files/uploads/papers/ESNTutorialRev.pdf>`__.
 #
 # You can observe outputs of teacher-forced networks that read with
 # coherent grammar but wander far from the correct translation -
 # intuitively it has learned to represent the output grammar and can "pick
 # up" the meaning once the teacher tells it the first few words, but it
 # has not properly learned how to create the sentence from the translation
 # in the first place.
 #
 # Because of the freedom PyTorch's autograd gives us, we can randomly
 # choose to use teacher forcing or not with a simple if statement. Turn
 # ``teacher_forcing_ratio`` up to use more of it.
 #

 teacher_forcing_ratio = 0.5

 # teacher forcing即指使用教师强迫其能够更快的收敛
 # 不过当训练好的网络被利用时，容易表现出不稳定性
 # teacher_forcing_ratio即指教师训练比率
 # 用于训练的函数

 def train(input_tensor, target_tensor, encoder, decoder, encoder_optimizer, decoder_optimizer, criterion,
           max_length=MAX_LENGTH):
     # encoder即指EncoderRNN(input_lang.n_words, hidden_size)
     # attn_decoder即指 AttnDecoderRNN(hidden_size, output_lang.n_words, dropout_p=0.1)
     # hidden=256
     encoder_hidden = encoder.initHidden()

     # encoder_optimizer 即指optim.SGD(encoder.parameters(), lr=learning_rate)
     # decoder_optimizer 即指optim.SGD(decoder.parameters(), lr=learning_rate)
     # nn.Parameter()是Variable的一种，常被用于模块参数(module parameter)。
     # Parameters 是 Variable 的子类。Paramenters和Modules一起使用的时候会有一些特殊的属性，
     # 即：当Paramenters赋值给Module的属性的时候，他会自动的被加到 Module的 参数列表中
     # (即：会出现在 parameters() 迭代器中)。将Varibale赋值给Module属性则不会有这样的影响。
     # 这样做的原因是：我们有时候会需要缓存一些临时的状态(state), 比如：模型中RNN的最后一个隐状态。
     # 如果没有Parameter这个类的话，那么这些临时变量也会注册成为模型变量。
     encoder_optimizer.zero_grad()
     decoder_optimizer.zero_grad()

     # 得到长度
     input_length = input_tensor.size(0)
     target_length = target_tensor.size(0)

     # 初始化outour值
     encoder_outputs = torch.zeros(max_length, encoder.hidden_size, device=device)

     loss = 0

     # 以下循环是学习过程
     for ei in range(input_length):
         encoder_output, encoder_hidden = encoder(input_tensor[ei], encoder_hidden)
         encoder_outputs[ei] = encoder_output[0, 0]  # 这里为什么取 0,0

     # 定义decoder的Input值
     decoder_input = torch.tensor([[SOS_token]], device=device)

     decoder_hidden = encoder_hidden

     use_teacher_forcing = True if random.random() < teacher_forcing_ratio else False

     if use_teacher_forcing:
         # Teacher forcing: Feed the target as the next input
         # 教师强制: 将目标作为下一个输入
         # 你观察教师强迫网络的输出,这些网络是用连贯的语法阅读的,但却远离了正确的翻译 -
         # 直观地来看它已经学会了代表输出语法,并且一旦老师告诉它前几个单词,就可以"拾取"它的意思,
         # 但它没有适当地学会如何从翻译中创建句子.
         for di in range(target_length):
             # 通过decoder得到输出值
             decoder_output, decoder_hidden, decoder_attention = decoder(
                 decoder_input, decoder_hidden, encoder_outputs)
             # 定义损失函数并计算
             loss += criterion(decoder_output, target_tensor[di])
             decoder_input = target_tensor[di]  # Teacher forcing

     else:
         # Without teacher forcing: use its own predictions as the next input
         # 没有教师强迫: 使用自己的预测作为下一个输入
         for di in range(target_length):
             # 通过decoder得到输出值
             decoder_output, decoder_hidden, decoder_attention = decoder(
                 decoder_input, decoder_hidden, encoder_outputs)

             # topk：第k个最小元素,返回第k个最小元素
             # 返回前k个最大元素，注意是前k个，largest=False，返回前k个最小元素
             # 此函数的功能是求取1-D 或N-D Tensor的最低维度的前k个最大的值，返回值为两个Tuple
             # 其中values是前k个最大值的Tuple，indices是对应的下标，默认返回结果是从大到小排序的。
             topv, topi = decoder_output.topk(1)
             decoder_input = topi.squeeze().detach()  # detach from history as input

             loss += criterion(decoder_output, target_tensor[di])
             if decoder_input.item() == EOS_token:
                 break
     # 反向传播
     loss.backward()

     # 更新参数
     encoder_optimizer.step()
     decoder_optimizer.step()

     return loss.item() / target_length

 ######################################################################
 # This is a helper function to print time elapsed and estimated time
 # remaining given the current time and progress %.
 #

 import time
 import math

 # 根据当前时间和进度百分比,这是一个帮助功能,用于打印经过的时间和估计的剩余时间.

 def asMinutes(s):
     m = math.floor(s / 60)
     s -= m * 60
     return '%dm %ds' % (m, s)

 def timeSince(since, percent):
     now = time.time()
     s = now - since
     es = s / (percent)
     rs = es - s
     return '%s (- %s)' % (asMinutes(s), asMinutes(rs))

 ######################################################################
 # The whole training process looks like this:
 #
 # -  Start a timer
 # -  Initialize optimizers and criterion
 # -  Create set of training pairs
 # -  Start empty losses array for plotting
 #
 # Then we call ``train`` many times and occasionally print the progress (%
 # of examples, time so far, estimated time) and average loss.
 #

 def trainIters(encoder, decoder, n_iters, print_every=1000, plot_every=100, learning_rate=0.01):
     start = time.time()
     plot_losses = []
     print_loss_total = 0  # Reset every print_every
     plot_loss_total = 0  # Reset every plot_every

     encoder_optimizer = optim.SGD(encoder.parameters(), lr=learning_rate)
     decoder_optimizer = optim.SGD(decoder.parameters(), lr=learning_rate)

     # 获取训练的一对样本
     training_pairs = [tensorsFromPair(random.choice(pairs))
                       for i in range(n_iters)]
     # 定义出的损失函数
     criterion = nn.NLLLoss()

     for iter in range(1, n_iters + 1):
         training_pair = training_pairs[iter - 1]
         input_tensor = training_pair[0]
         target_tensor = training_pair[1]

         # 训练的过程并用于当损失函数
         loss = train(input_tensor, target_tensor, encoder,
                      decoder, encoder_optimizer, decoder_optimizer, criterion)
         print_loss_total += loss
         plot_loss_total += loss

         if iter % print_every == 0:
             print_loss_avg = print_loss_total / print_every
             print_loss_total = 0
             # 打印进度(样本的百分比,到目前为止的时间,估计的时间)和平均损失.
             print('%s (%d %d%%) %.4f' % (timeSince(start, iter / n_iters),
                                          iter, iter / n_iters * 100, print_loss_avg))

         if iter % plot_every == 0:
             plot_loss_avg = plot_loss_total / plot_every
             plot_losses.append(plot_loss_avg)
             plot_loss_total = 0
     # 绘制图像
     showPlot(plot_losses)

 ######################################################################
 # Plotting results
 # ----------------
 #
 # Plotting is done with matplotlib, using the array of loss values
 # ``plot_losses`` saved while training.
 #

 import matplotlib.pyplot as plt

 plt.switch_backend('agg')
 import matplotlib.ticker as ticker
 import numpy as np

 # 使用matplotlib进行绘图，使用训练时保存的损失值plot_losses数组.
 def showPlot(points):
     plt.figure()
     fig, ax = plt.subplots()
     # this locator puts ticks at regular intervals
     # 这个定位器会定期发出提示信息
     loc = ticker.MultipleLocator(base=0.2)
     ax.yaxis.set_major_locator(loc)
     plt.plot(points)

 ######################################################################
 # Evaluation
 # ==========
 #
 # Evaluation is mostly the same as training, but there are no targets so
 # we simply feed the decoder's predictions back to itself for each step.
 # Every time it predicts a word we add it to the output string, and if it
 # predicts the EOS token we stop there. We also store the decoder's
 # attention outputs for display later.
 #

 def evaluate(encoder, decoder, sentence, max_length=MAX_LENGTH):
     with torch.no_grad():
         # 从sentence中得到对应的变量
         input_tensor = tensorFromSentence(input_lang, sentence)
         # 长度
         input_length = input_tensor.size()[0]

         # encoder即指EncoderRNN(input_lang.n_words, hidden_size)
         # attn_decoder即指 AttnDecoderRNN(hidden_size,
         # output_lang.n_words, dropout_p=0.1)
         # hidden=256
         encoder_hidden = encoder.initHidden()

         # 初始化outputs值
         encoder_outputs = torch.zeros(max_length, encoder.hidden_size, device=device)

         # 以下是学习过程
         for ei in range(input_length):
             encoder_output, encoder_hidden = encoder(input_tensor[ei],
                                                      encoder_hidden)
             encoder_outputs[ei] += encoder_output[0, 0]

         # 定义好decoder部分的input值
         decoder_input = torch.tensor([[SOS_token]], device=device)  # SOS

         # 设置好隐藏层
         decoder_hidden = encoder_hidden

         decoded_words = []
         decoder_attentions = torch.zeros(max_length, max_length)

         for di in range(max_length):
             # 得到结果
             decoder_output, decoder_hidden, decoder_attention = decoder(decoder_input, decoder_hidden, encoder_outputs)

             # attention部分的数据
             decoder_attentions[di] = decoder_attention.data
             # 选择output中的第一个值
             topv, topi = decoder_output.data.topk(1)
             if topi.item() == EOS_token:
                 decoded_words.append('<EOS>')
                 break
             else:
                 decoded_words.append(output_lang.index2word[topi.item()])  # 将output_lang添加到decoded

             decoder_input = topi.squeeze().detach()

         return decoded_words, decoder_attentions[:di + 1]

 ######################################################################
 # We can evaluate random sentences from the training set and print out the
 # input, target, and output to make some subjective quality judgements:
 #

 # 从训练集中评估随机的句子并打印出输入,目标和输出以作出一些主观质量判断
 def evaluateRandomly(encoder, decoder, n=10):
     for i in range(n):
         pair = random.choice(pairs)
         print('>', pair[0])
         print('=', pair[1])
         output_words, attentions = evaluate(encoder, decoder, pair[0])
         output_sentence = ' '.join(output_words)
         print('<', output_sentence)
         print('')

 ######################################################################
 # Training and Evaluating
 # =======================
 #
 # With all these helper functions in place (it looks like extra work, but
 # it makes it easier to run multiple experiments) we can actually
 # initialize a network and start training.
 #
 # Remember that the input sentences were heavily filtered. For this small
 # dataset we can use relatively small networks of 256 hidden nodes and a
 # single GRU layer. After about 40 minutes on a MacBook CPU we'll get some
 # reasonable results.
 #
 # .. Note::
 #    If you run this notebook you can train, interrupt the kernel,
 #    evaluate, and continue training later. Comment out the lines where the
 #    encoder and decoder are initialized and run ``trainIters`` again.
 #

 hidden_size = 256
 # 编码部分
 encoder1 = EncoderRNN(input_lang.n_words, hidden_size).to(device)
 # 加入了attention机制的解码部分
 attn_decoder1 = AttnDecoderRNN(hidden_size, output_lang.n_words, dropout_p=0.1).to(device)
 # 训练部分
 trainIters(encoder1, attn_decoder1, 75000, print_every=5000)

 ######################################################################
 # 随机生成一组结果
 evaluateRandomly(encoder1, attn_decoder1)

 ######################################################################
 # Visualizing Attention
 # ---------------------
 #
 # A useful property of the attention mechanism is its highly interpretable
 # outputs. Because it is used to weight specific encoder outputs of the
 # input sequence, we can imagine looking where the network is focused most
 # at each time step.
 #
 # You could simply run ``plt.matshow(attentions)`` to see attention output
 # displayed as a matrix, with the columns being input steps and rows being
 # output steps:
 #

 output_words, attentions = evaluate(encoder1, attn_decoder1, "je suis trop froid .")
 plt.matshow(attentions.numpy())

 ######################################################################
 # For a better viewing experience we will do the extra work of adding axes
 # and labels:

 def showAttention(input_sentence, output_words, attentions):
     # Set up figure with colorbar
     fig = plt.figure()
     ax = fig.add_subplot(111)
     cax = ax.matshow(attentions.numpy(), cmap='bone')
     fig.colorbar(cax)

     # Set up axes
     ax.set_xticklabels([''] + input_sentence.split(' ') +
                        ['<EOS>'], rotation=90)
     ax.set_yticklabels([''] + output_words)

     # Show label at every tick
     ax.xaxis.set_major_locator(ticker.MultipleLocator(1))
     ax.yaxis.set_major_locator(ticker.MultipleLocator(1))

     plt.show()

 def evaluateAndShowAttention(input_sentence):
     output_words, attentions = evaluate(
         encoder1, attn_decoder1, input_sentence)
     print('input =', input_sentence)
     print('output =', ' '.join(output_words))
     showAttention(input_sentence, output_words, attentions)

 evaluateAndShowAttention("elle a cinq ans de moins que moi .")
 evaluateAndShowAttention("elle est trop petit .")
 evaluateAndShowAttention("je ne crains pas de mourir .")
 evaluateAndShowAttention("c est un jeune directeur plein de talent .")

 ######################################################################
 # Exercises
 # =========
 #
 # -  Try with a different dataset
 #
 #    -  Another language pair
 #    -  Human → Machine (e.g. IOT commands)
 #    -  Chat → Response
 #    -  Question → Answer
 #
 # -  Replace the embeddings with pre-trained word embeddings such as word2vec or
 #    GloVe
 # -  Try with more layers, more hidden units, and more sentences. Compare
 #    the training time and results.
 # -  If you use a translation file where pairs have two of the same phrase
 #    (``I am test \t I am test``), you can use this as an autoencoder. Try
 #    this:
 #
 #    -  Train as an autoencoder
 #    -  Save only the Encoder network
 #    -  Train a new Decoder for translation from there
 #