NLP实战7：seq2seq翻译实战-Pytorch复现

本文为[365天深度学习训练营]内部限免文章（版权归 *K同学啊* 所有）
作者：[K同学啊]

本周任务：
●请根据N5、N6周内容，为解码器添加上注意力机制

一、前期准备

准备好数据集eng-fra.txt

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)

cuda

1. 搭建语言类

SOS_token = 0
EOS_token = 1

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)

    def addWord(self, word):
        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
        else:
            self.word2count[word] += 1

2. 文本处理函数

def unicodeToAscii(s):
    return ''.join(
        c for c in unicodedata.normalize('NFD', s)
        if unicodedata.category(c) != 'Mn'
    )

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

3. 文件读取函数

def readLangs(lang1, lang2, reverse=False):
    print("Reading lines...")

    # 以行为单位读取文件
    lines = open('%s-%s.txt'%(lang1,lang2), encoding='utf-8').\
            read().strip().split('\n')

    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:
        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

.startswith(eng_prefixes) 是字符串方法 startswith() 的调用。它用于检查一个字符串是否以指定的前缀开始。

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))

 

二、Seq2Seq 模型

2.1 编码器（Encoder）

class EncoderRNN(nn.Module):
    def __init__(self, input_size, hidden_size):
        super(EncoderRNN, self).__init__()
        self.hidden_size = hidden_size
        self.embedding   = nn.Embedding(input_size, hidden_size)
        self.gru         = nn.GRU(hidden_size, hidden_size)

    def forward(self, input, hidden):
        embedded       = self.embedding(input).view(1, 1, -1)
        output         = embedded
        output, hidden = self.gru(output, hidden)
        return output, hidden

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

2.2 解码器（Decoder）

这里的代码与N6的不同，加入了注意力机制。

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):
        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)

三、训练

3.1 数据预处理

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

 在序列生成任务（如机器翻译或文本生成）中，解码器的输入通常是由解码器自己生成的预测结果。然而，这种自回归方式可能导致累积误差和输出偏离。为解决此问题，引入了"Teacher Forcing"技术，将目标序列的真实值作为解码器的输入，提供更准确的指导信号。在代码中，通过use_teacher_forcing变量选择策略：当为True时，采用"Teacher Forcing"，使用真实标签作为输入；当为False时，采用"Without Teacher Forcing"，使用解码器的预测结果作为输入。使用use_teacher_forcing可以平衡预测能力和稳定性。Teacher Forcing可以加速训练收敛，而Without Teacher Forcing更适合真实场景。一般会逐渐减小Teacher Forcing的比例，以使模型过渡到更自主的生成模式。在代码中，.topk(1)获取decoder_output中最大元素和索引，.squeeze().detach()对topi进行处理，作为下一个解码器的输入。

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('')
                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('')

四、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, 20000, print_every=5000)

 

evaluateRandomly(encoder1, attn_decoder1)

> il est dans la lune .
= he is distracted .
< he s in the . .

> vous etes tres astucieux .
= you re very astute .
< you re very timid .

> vous etes tenace .
= you re resilient .
< you re clever .

> il est dans le petrin .
= he is in trouble now .
< he is in the .

> je vais bientot etre partie .
= i m going to be gone soon .
< i m going to go . . .

1. Loss图

import matplotlib.pyplot as plt
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()

2. 可视化注意力

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

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(' ') +
                       [''], 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 is years years older than i am .

 input = elle est trop petit .
output = she is too drunk .

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

input = c est un jeune directeur plein de talent .
output = he s a talented young young .