第14周-Seq2Seq模型-NLP

• 🍨 本文为🔗365天深度学习训练营 中的学习记录博客
• *🍖 参考：K同学啊

1.导入所需库和函数

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")
print(device)

cpu

1.1搭建语言类

SOS_token=0
EOS_token=1
#语言类，方便对语料库进行操作
class Lang:
    def __init__(self,name):
        # 初始化语言类，传入语言的名称
        self.name=name
        # 初始化单词到索引的映射字典
        self.word2index={}
        # 初始化单词到出现次数的映射字典
        self.word2count={}
        # 初始化索引到单词的映射字典，包含SOS（开始标记）和EOS（结束标记）
        self.index2word={0:"SOS",1:"EOS"}
        self.n_words=2 #初始化为2，因为SOS和EOS已经存在
    def addSentence(self,sentence):#将句子中的每个单词添加到语言类中
        for word in sentence.split(' '):
            self.addWord(word)
    def addWord(self,word):#将单词添加到语言类中
        if word not in self.word2index:
            self.word2index[word]=self.n_words
            self.word2count[word]=1
            self.index2word[self.n_words]=word
            self.n_words += 1  # **确保索引值递增**
        else:
            self.word2count[word]+=1

1.2文本读取函数

def unicodeToAscii(s):#将unicode编码的字符串转换为ascii编码的字符串
    return ''.join(
        c for c in unicodedata.normalize('NFD', s)
        if unicodedata.category(c) != 'Mn'
    )
def normalizeString(s):#将字符串中的标点符号转换为小写，并去除特殊字符
    s=unicodeToAscii(s.lower().strip())
    s=re.sub(r"([.!?])",r" \1",s)
    s=re.sub(r"[^a-zA-Z.!?]+",r" ",s)
    return s

1.2文件处理函数

def unicodeToAscii(s):
    # 使用unicodedata库的normalize函数将输入字符串s进行Unicode规范化
    # 'NFD'表示Unicode规范化形式D，即将组合字符分解为基本字符和组合字符
    # 例如，将带重音符号的字符分解为基本字符和单独的重音符号
    return ''.join(
        c for c in unicodedata.normalize('NFD', s)
        # 遍历规范化后的字符，使用生成器表达式筛选出不需要的字符
        # unicodedata.category(c)返回字符c的Unicode类别
        # 'Mn'表示Mark, Nonspacing，即非间距标记，通常用于重音符号
        if unicodedata.category(c) != 'Mn'
        # 只保留非'Mn'类别的字符，即去除重音符号
    )

# 小写化，剔除标点与非字母符号
def normalizeString(s):
    # 将输入字符串转换为小写并去除首尾的空白字符
    s = unicodeToAscii(s.lower().strip())
    # 在标点符号前后添加空格，使其更加标准化
    s = re.sub(r"([.!?])", r" \1", s)
    # 移除所有非字母和标点符号的字符，只保留字母和标点符号，并用空格替换其他字符
    s = re.sub(r"[^a-zA-Z.!?]+", r" ", s)
    # 返回处理后的字符串
    return s

1.3文件读取函数

def readLangs(lang1, lang2, reverse=False):
    # 打印提示信息，表示开始读取文件
    print("Reading lines...")

    # 以行为单位读取文件
    # 使用字符串格式化将语言代码插入文件路径
    # 使用utf-8编码读取文件，并去除首尾空白字符，然后按行分割
    lines = open('%s-%s.txt'%(lang1,lang2), encoding='utf-8').\
            read().strip().split('\n')

    # 将每一行放入一个列表中
    # 一个列表中有两个元素，A语言文本与B语言文本
    # 使用列表推导式遍历每一行，并使用split('\t')分割每一行
    # 对分割后的字符串进行标准化处理
    pairs = [[normalizeString(s) for s in l.split('\t')] for l in lines]

    # 创建Lang实例，并确认是否反转语言顺序
    if reverse:
        # 如果reverse为True，反转语言对
        pairs       = [list(reversed(p)) for p in pairs]
        # 创建输入语言实例
        input_lang  = Lang(lang2)
        # 创建输出语言实例
        output_lang = Lang(lang1)
    else:
        # 如果reverse为False，保持语言对顺序
        # 创建输入语言实例
        input_lang  = Lang(lang1)
        # 创建输出语言实例
        output_lang = Lang(lang2)

    # 返回输入语言实例、输出语言实例和语言对列表
    return input_lang, output_lang, pairs

MAX_LENGTH = 10      # 定义语料最长长度

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)

def filterPairs(pairs):
    # 选取仅仅包含 eng_prefixes 开头的语料
    return [pair for pair in pairs if filterPair(pair)]

def prepareData(lang1, lang2, reverse=False):
    # 读取文件中的数据
    input_lang, output_lang, pairs = readLangs(lang1, lang2, reverse)
    print("Read %s sentence pairs" % len(pairs))
    
    # 按条件选取语料
    pairs = filterPairs(pairs[:])
    print("Trimmed to %s sentence pairs" % len(pairs))
    print("Counting words...")
    
    # 将语料保存至相应的语言类
    for pair in pairs:
        input_lang.addSentence(pair[0])
        output_lang.addSentence(pair[1])
        
    # 打印语言类的信息    
    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

input_lang, output_lang, pairs = prepareData('eng', 'fra', True)
print(random.choice(pairs))

Reading lines...
Read 135842 sentence pairs
Trimmed to 10599 sentence pairs
Counting words...
Counted words:
fra 4345
eng 2803
['je crains que ce soit impossible .', 'i m afraid that s impossible .']

2.Seq2Seq模型

2.1编码器

class EncoderRNN(nn.Module):
    # 定义一个名为EncoderRNN的类，继承自nn.Module，用于实现编码器RNN
    def __init__(self, input_size, hidden_size):
        # 构造函数，初始化EncoderRNN对象
        super(EncoderRNN, self).__init__()
        # 调用父类nn.Module的构造函数
        self.hidden_size = hidden_size
        # 将隐藏层大小保存为对象的属性
        self.embedding   = nn.Embedding(input_size, hidden_size)
        # 创建一个嵌入层，将输入的单词索引映射到隐藏层大小的向量空间
        self.gru         = nn.GRU(hidden_size, hidden_size)

        # 创建一个GRU（门控循环单元）层，输入和输出维度均为hidden_size
    def forward(self, input, hidden):
        # 定义前向传播函数，接收输入和隐藏状态
        embedded       = self.embedding(input).view(1, 1, -1)
        # 将输入通过嵌入层转换为向量，并调整形状为(1, 1, hidden_size)
        output         = embedded
        # 将嵌入向量作为初始输出
        output, hidden = self.gru(output, hidden)
        # 将输出和隐藏状态传入GRU层，得到新的输出和隐藏状态
        return output, hidden

        # 返回新的输出和隐藏状态
    def initHidden(self):
        # 定义初始化隐藏状态的函数
        return torch.zeros(1, 1, self.hidden_size, device=device)

2.2解码器

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):
        # 将输入的单词索引通过嵌入层转换为词向量，并调整形状为(1, 1, -1)
        embedded = self.embedding(input).view(1, 1, -1)
        embedded = self.dropout(embedded)

        attn_weights = F.softmax(
            self.attn(torch.cat((embedded[0], hidden[0]), 1)), dim=1)
        attn_applied = torch.bmm(attn_weights.unsqueeze(0),
                                 encoder_outputs.unsqueeze(0))

        output = torch.cat((embedded[0], attn_applied[0]), 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)

3.训练

3.1数据预处理

# 将文本数字化，获取词汇index
def indexesFromSentence(lang, sentence):
    return [lang.word2index[word] for word in sentence.split(' ')]

# 将数字化的文本，转化为tensor数据
def tensorFromSentence(lang, sentence):
    indexes = indexesFromSentence(lang, sentence)
    indexes.append(EOS_token)
    return torch.tensor(indexes, dtype=torch.long, device=device).view(-1, 1)

# 输入pair文本，输出预处理好的数据
def tensorsFromPair(pair):
    input_tensor  = tensorFromSentence(input_lang, pair[0])
    target_tensor = tensorFromSentence(output_lang, pair[1])
    return (input_tensor, target_tensor)

3.2训练函数

teacher_forcing_ratio = 0.5

def train(input_tensor, target_tensor, 
          encoder, decoder, 
          encoder_optimizer, decoder_optimizer, 
          criterion, max_length=MAX_LENGTH):
    
    # 编码器初始化
    encoder_hidden = encoder.initHidden()
    
    # grad属性归零
    encoder_optimizer.zero_grad()
    decoder_optimizer.zero_grad()

    input_length  = input_tensor.size(0)
    target_length = target_tensor.size(0)
    
    # 用于创建一个指定大小的全零张量（tensor），用作默认编码器输出
    encoder_outputs = torch.zeros(max_length, encoder.hidden_size, device=device)

    loss = 0
    
    # 将处理好的语料送入编码器
    for ei in range(input_length):
        encoder_output, encoder_hidden = encoder(input_tensor[ei], encoder_hidden)
        encoder_outputs[ei]            = encoder_output[0, 0]
    
    # 解码器默认输出
    decoder_input  = torch.tensor([[SOS_token]], device=device)
    decoder_hidden = encoder_hidden

    use_teacher_forcing = True if random.random() < teacher_forcing_ratio else False
    
    # 将编码器处理好的输出送入解码器
    if use_teacher_forcing:
        # Teacher forcing: Feed the target as the next input
        for di in range(target_length):
            decoder_output, decoder_hidden, decoder_attention = decoder(
                decoder_input, decoder_hidden, encoder_outputs)
            
            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
        for di in range(target_length):
            decoder_output, decoder_hidden, decoder_attention = decoder(
                decoder_input, decoder_hidden, encoder_outputs)
            
            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

    loss.backward()

    encoder_optimizer.step()
    decoder_optimizer.step()

    return loss.item() / target_length

import time
import math

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))

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)
    
    # 在 pairs 中随机选取 n_iters 条数据用作训练集
    training_pairs    = [tensorsFromPair(random.choice(pairs)) for i in range(n_iters)]
    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]

        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
            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

    return plot_losses

3.3评估

def evaluate(encoder, decoder, sentence, max_length=MAX_LENGTH):
    with torch.no_grad():
        input_tensor    = tensorFromSentence(input_lang, sentence)
        input_length    = input_tensor.size()[0]
        encoder_hidden  = encoder.initHidden()

        encoder_outputs = torch.zeros(max_length, encoder.hidden_size, device=device)

        for ei in range(input_length):
            encoder_output, encoder_hidden = encoder(input_tensor[ei],encoder_hidden)
            encoder_outputs[ei]           += encoder_output[0, 0]

        decoder_input  = torch.tensor([[SOS_token]], device=device)  # SOS

        decoder_hidden = encoder_hidden

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

        for di in range(max_length):
            decoder_output, decoder_hidden, decoder_attention = decoder(
                decoder_input, decoder_hidden, encoder_outputs)
            
            decoder_attentions[di] = decoder_attention.data
            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()])

            decoder_input = topi.squeeze().detach()

        return decoded_words, decoder_attentions[:di + 1]

def evaluateRandomly(encoder, decoder, n=5):
    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('')

4.训练与评估

hidden_size   = 256
encoder1      = EncoderRNN(input_lang.n_words, hidden_size).to(device)
attn_decoder1 = AttnDecoderRNN(hidden_size, output_lang.n_words, dropout_p=0.1).to(device)

plot_losses   = trainIters(encoder1, attn_decoder1, 10000, print_every=5000)

2m 25s (- 2m 25s) (5000 50%) 2.8408
4m 52s (- 0m 0s) (10000 100%) 2.3021

evaluateRandomly(encoder1, attn_decoder1)

> il est toujours ponctuel pour ses rendez vous .
= he s always on time for his appointments .
< he s always with the for . <EOS>

> nous sommes pas en train de regarder .
= we re not looking .
< we re not . . . <EOS>

> vous n allez pas le croire .
= you re not going to believe this .
< you re not the . . . <EOS>

> tu es degoutante .
= you re disgusting .
< you re crafty . <EOS>

> vous etes seule n est ce pas ?
= you re alone aren t you ?
< you re aren aren t he ? <EOS>

​

4.1loss图

import matplotlib.pyplot as plt
#隐藏警告
import warnings
warnings.filterwarnings("ignore")               # 忽略警告信息
# plt.rcParams['font.sans-serif']    = ['SimHei'] # 用来正常显示中文标签
plt.rcParams['axes.unicode_minus'] = False      # 用来正常显示负号
plt.rcParams['figure.dpi']         = 100        # 分辨率

epochs_range = range(len(plot_losses))

plt.figure(figsize=(8, 3))

plt.subplot(1, 1, 1)
plt.plot(epochs_range, plot_losses, label='Training Loss')
plt.legend(loc='upper right')
plt.title('Training Loss')
plt.show()

​

4.2可视化注意力

import matplotlib.pyplot as plt

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

<matplotlib.image.AxesImage at 0x1ad9751c1f0>

​

​

import matplotlib.ticker as ticker
#隐藏警告
import warnings
warnings.filterwarnings("ignore")               # 忽略警告信息

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 .")

input = elle a cinq ans de moins que moi .
output = she s looking of her me . <EOS>

input = elle est trop petit .
output = she s too tired . <EOS>

input = je ne crains pas de mourir .
output = i m not afraid of <EOS>

input = c est un jeune directeur plein de talent .
output = he is my friend . <EOS>