前言
神经网络机器翻译(NMT, neuro machine tranlation)是AIGC发展道路上的一个重要应用。正是对这个应用的研究，发展出了注意力机制，在此基础上产生了AIGC领域的霸主transformer。我们今天先把注意力机制这些东西放一边，介绍一个对机器翻译起到重要里程碑作用的模型：LSTM encoder-decoder模型（sutskever et al. 2014）。根据这篇文章的描述，这个模型不需要特别的优化，就可以取得超过其他NMT模型的效果，所以我们也来动手实现一下，看看是不是真的有这么厉害。
模型
原文作者采用了4层LSTM模型，每层有1000个单元（每个单元有输入门，输出门，遗忘门和细胞状态更新共计4组状态），采用1000维单词向量，纯RNN部分，就有64M参数。同时，在encoder的输出，和decoder的输出后放一个长度为80000的softmax层（因为论文的输出字典长80000），用于softmax的参数量为320M。整个模型共计320M + 64M = 384M。该模型用了8GPU的服务器训练了10天。
模型大概长这样：
按照现在的算力价格，用8张4090的主机训练每小时要花20多块钱，训练一轮下来需要花费小5000，笔者当然没有这么土豪，所以我们会使用一个参数量小得多的模型，主要为了记录整个搭建过程使用到的工具链和技术。另外，由于笔者使用了一个预训练的词向量库，包含了中英文单词共计128万多条，其中中文90多万，英文30多万，要像论文中一样用一个超大的softmax来预测每个词的概率并不现实，因此先使用一个linear层再加上relu来简化，加快训练过程，只求能看到收敛。
笔者的模型看起来像这样：
该模型的主要参数如下：
词向量维度：300
LSTM隐藏层个数：600
LSTM层数：4
linear层输入：600
linear层输出：300
模型参数个数如下为：
==========================================================================================
Layer (type:depth-idx) Output Shape Param #
==========================================================================================
Seq2Seq [1, 11, 300] --
├─Encoder: 1-1 [1, 300] --
│ └─LSTM: 2-1 [1, 10, 600] 10,819,200
│ └─Linear: 2-2 [1, 300] 180,300
│ └─ReLU: 2-3 [1, 300] --
├─Decoder: 1-2 [1, 11, 300] --
│ └─LSTM: 2-4 [1, 11, 600] 10,819,200
│ └─Linear: 2-5 [1, 11, 300] 180,300
│ └─ReLU: 2-6 [1, 11, 300] --
==========================================================================================
Total params: 21,999,000
Trainable params: 21,999,000
Non-trainable params: 0
Total mult-adds (M): 227.56
==========================================================================================
Input size (MB): 0.02
Forward/backward pass size (MB): 0.13
Params size (MB): 88.00
Estimated Total Size (MB): 88.15
==========================================================================================
如果大家希望了解LSTM层的10,819,200个参数如何计算出来，可以参考pytorch源码 pytorch/torch/csrc/api/src/nn/modules/rnn.cpp中方法void RNNImplBase::reset()的实现。笔者如果日后有空也可能会写一写。
3 单词向量及语料
3.1 语料
先说语料，NMT需要大量的平行语料，语料可以从这里获取。另外有个语料天涯网站大全分享给大家。
3.2 词向量
首先需要对句子进行分词，中英文都需要做分词。中文分词工具本例采用jieba，可直接安装。
$ pip install jieba
...
$ python
Python 3.11.6 (tags/v3.11.6:8b6ee5b, Oct 2 2023, 14:57:12) [MSC v.1935 64 bit (AMD64)] on win32
Type "help", "copyright", "credits" or "license" for more information.
>>> for token in jieba.cut("我爱踢足球！", cut_all=True):
... print(token)
...
我
爱
踢足球
足球
！
英文分词采用nltk，安装之后，需要下载一个分词模型。
$ pip install nltk
...
$ python
Python 3.11.6 (tags/v3.11.6:8b6ee5b, Oct 2 2023, 14:57:12) [MSC v.1935 64 bit (AMD64)] on win32
Type "help", "copyright", "credits" or "license" for more information.
>>> import nltk
>>> nltk.download("punkt")
...
>>> from nltk import word_tokenize
>>> word_tokenize('i love you')
['i', 'love', 'you']
国内有墙，一般下载不了，所以可以到这里找到punkt文件并下载，解压到~/nltk_data/tokenizers/下边。
3.3 加载语料代码
import xml.etree.ElementTree as ETclass TmxHandler():def __init__(self):self.tag=Noneself.lang=Noneself.corpus={}def handleStartTu(self, tag):self.tag=tagself.lang=Noneself.corpus={}def handleStartTuv(self, tag, attributes):if self.tag == 'tu':if attributes['{http://www.w3.org/XML/1998/namespace}lang']:self.lang=attributes['{http://www.w3.org/XML/1998/namespace}lang']else:raise Exception('tuv element must has a xml:lang attribute')self.tag = tagelse:raise Exception('tuv element must go under tu, not ' + tag)def handleStartSeg(self, tag, elem):if self.tag == 'tuv':self.tag = tagif self.lang:if elem.text:self.corpus[self.lang]=elem.textelse:raise Exception('lang must not be none')else:raise Exception('seg element must go under tuv, not ' + tag)def startElement(self, tag, attributes, elem):if tag== 'tu':self.handleStartTu(tag)elif tag == 'tuv':self.handleStartTuv(tag, attributes)elif tag == 'seg':self.handleStartSeg(tag, elem)def endElem(self, tag):if self.tag and self.tag != tag:raise Exception(self.tag + ' could not end with ' + tag)if tag == 'tu':self.tag=Noneself.lang=Noneself.corpus={}elif tag == 'tuv':self.tag='tu'self.lang=Noneelif tag == 'seg':self.tag='tuv'def parse(self, filename):for event, elem in ET.iterparse(filename, events=('start','end')):if event == 'start':self.startElement(elem.tag, elem.attrib, elem)elif event == 'end':if elem.tag=='tu':yield self.corpusself.endElem(elem.tag)
3.4 句子转词向量代码
from gensim.models import KeyedVectors
import torch
import jieba
from nltk import word_tokenize
import numpy as npclass WordEmbeddingLoader():def __init__(self):passdef load(self, fname):self.model = KeyedVectors.load_word2vec_format(fname)def get_embeddings(self, word:str):if self.model:try:return self.model.get_vector(word)except(KeyError):return Noneelse:return Nonedef get_scentence_embeddings(self, scent:str, lang:str):embeddings = []ws = []if(lang == 'zh'):ws = jieba.cut(scent, cut_all=True)elif lang == 'en':ws = word_tokenize(scent)else:raise Exception('Unsupported language ' + lang)for w in ws:embedding = self.get_embeddings(w.lower())if embedding is None:embedding = np.zeros(self.model.vector_size)embedding = torch.from_numpy(embedding).float()embeddings.append(embedding.unsqueeze(0))return torch.cat(embeddings, dim=0)
4 模型代码实现
4.1 encoder
import torch.nn as nnclass Encoder(nn.Module):def __init__(self, device, embeddings=300, hidden_size=600, num_layers=4):super().__init__()self.device = deviceself.hidden_layer_size = hidden_sizeself.n_layers = num_layersself.embedding_size = embeddingsself.lstm = nn.LSTM(embeddings, hidden_size, num_layers, batch_first=True)self.linear = nn.Linear(hidden_size, embeddings)self.relu = nn.ReLU()def forward(self, x):# x: [batch size, seq length, embeddings]# lstm_out: [batch size, x length, hidden size]lstm_out, (hidden, cell) = self.lstm(x)# linear input is the lstm output of the last wordlineared = self.linear(lstm_out[:,-1,:].squeeze(1))out = self.relu(lineared)# hidden: [n_layer, batch size, hidden size]# cell: [n_layer, batch size, hidden size]return out, hidden, cell
4.2 decoder
import torch.nn as nnclass Decoder(nn.Module):def __init__(self, device, embedding_size=300, hidden_size=900, num_layers=4):super().__init__()self.device = deviceself.hidden_layer_size = hidden_sizeself.n_layers = num_layersself.embedding_size = embedding_sizeself.lstm = nn.LSTM(embedding_size, hidden_size, num_layers, batch_first=True)self.linear = nn.Linear(hidden_size, embedding_size)self.relu = nn.ReLU()def forward(self, x, hidden_in, cell_in):# x: [batch_size, x length, embeddings]# hidden: [n_layers, batch size, hidden size]# cell: [n_layers, batch size, hidden size]# lstm_out: [seq length, batch size, hidden size]lstm_out, (hidden,cell) = self.lstm(x, (hidden_in, cell_in))# prediction: [seq length, batch size, embeddings]prediction=self.relu(self.linear(lstm_out))return prediction, hidden, cell
4.3 encoder-decoder
接下来把encoder和decoder串联起来。
import torch
import encoder as enc
import decoder as dec
import torch.nn as nn
import timeclass Seq2Seq(nn.Module):def __init__(self, device, embeddings, hiddens, n_layers):super().__init__()self.device = deviceself.encoder = enc.Encoder(device, embeddings, hiddens, n_layers)self.decoder= dec.Decoder(device, embeddings, hiddens, n_layers)self.embeddings = self.encoder.embedding_sizeassert self.encoder.n_layers == self.decoder.n_layers, "Number of layers of encoder and decoder must be equal!"assert self.decoder.hidden_layer_size==self.decoder.hidden_layer_size, "Hidden layer size of encoder and decoder must be equal!"# x: [batches, x length, embeddings]# x is the source scentences# y: [batches, y length, embeddings]# y is the target scentencesdef forward(self, x, y):# encoder_out: [batches, n_layers, embeddings]# hidden, cell: [n layers, batch size, embeddings]encoder_out, hidden, cell = self.encoder(x)# use encoder output as the first word of the decode sequencedecoder_input = torch.cat((encoder_out.unsqueeze(0), y), dim=1)# predicted: [batches, y length, embeddings]predicted, hidden, cell = self.decoder(decoder_input, hidden, cell)return predicted
5 模型训练
5.1 训练代码
def do_train(model:Seq2Seq, train_set, optimizer, loss_function):step = 0model.train()# seq: [seq length, embeddings]# labels: [label length, embeddings]for seq, labels in train_set:step = step + 1# ignore the last word of the label scentence# because it is to be predictedlabel_input = labels[:-1].unsqueeze(0)# seq_input: [1, seq length, embeddings]seq_input = seq.unsqueeze(0)# y_pred: [1, seq length + 1, embeddings]y_pred = model(seq_input, label_input)# single_loss = loss_function(y_pred.squeeze(0), labels.to(self.device))single_loss = loss_function(y_pred.squeeze(0), labels)optimizer.zero_grad()single_loss.backward()optimizer.step()print_steps = 100if print_steps != 0 and step%print_steps==1:print(f'[step: {step} - {time.asctime(time.localtime(time.time()))}] - loss:{single_loss.item():10.8f}')def train(device, model, embedding_loader, corpus_fname, batch_size:int, batches: int):reader = corpus_reader.TmxHandler()loss = torch.nn.MSELoss()# summary(model, input_size=[(1, 10, 300),(1,10,300)])optimizer = torch.optim.SGD(model.parameters(), lr=0.01)generator = reader.parse(corpus_fname)for _b in range(batches):batch = []try:for _c in range(batch_size):try:corpus = next(generator)if 'en' in corpus and 'zh' in corpus:en = embedding_loader.get_scentence_embeddings(corpus['en'], 'en').to(device)zh = embedding_loader.get_scentence_embeddings(corpus['zh'], 'zh').to(device)batch.append((en,zh))except (StopIteration):breakfinally:print(time.localtime())print("batch: " + str(_b))do_train(model, batch, optimizer, loss)torch.save(model, "./models/seq2seq_" + str(time.time()))if __name__=="__main__":# device = torch.device('cuda')device = torch.device('cpu')embeddings = 300hiddens = 600n_layers = 4embedding_loader = word2vec.WordEmbeddingLoader()print("loading embedding")# a full vocabulary takes too long to load, a baby vocabulary is used for demo purposeembedding_loader.load("../sgns.merge.word.toy")print("load embedding finished")# if there is an existing model, load the existing model from file# model_fname = "./models/_seq2seq_1698000846.3281412"model_fname = Nonemodel = Noneif not model_fname is None:print('loading model from ' + model_fname)model = torch.load(model_fname, map_location=device)print('model loaded')else:model = Seq2Seq(device, embeddings, hiddens, n_layers).to(device)train(device, model, embedding_loader, "../News-Commentary_v16.tmx", 1000, 100)
5.2 使用CPU进行训练
让我们先来体验一下CPU的龟速训练。下图是每100句话的训练输出。每次打印的间隔大约为2-3分钟。
[step: 1 - Thu Oct 26 05:14:13 2023] - loss:0.00952744
[step: 101 - Thu Oct 26 05:17:11 2023] - loss:0.00855174
[step: 201 - Thu Oct 26 05:20:07 2023] - loss:0.00831730
[step: 301 - Thu Oct 26 05:23:09 2023] - loss:0.00032693
[step: 401 - Thu Oct 26 05:25:55 2023] - loss:0.00907284
[step: 501 - Thu Oct 26 05:28:55 2023] - loss:0.00937218
[step: 601 - Thu Oct 26 05:32:00 2023] - loss:0.00823146
5.3 使用GPU进行训练
如果把main函数的第一行中的"cpu"改成“cuda”，则可以使用显卡进行训练。笔者使用的是一张GTX1660显卡，打印间隔缩短为15秒。
[step: 1 - Thu Oct 26 06:38:45 2023] - loss:0.00955237
[step: 101 - Thu Oct 26 06:38:50 2023] - loss:0.00844441
[step: 201 - Thu Oct 26 06:38:56 2023] - loss:0.00820994
[step: 301 - Thu Oct 26 06:39:01 2023] - loss:0.00030389
[step: 401 - Thu Oct 26 06:39:06 2023] - loss:0.00896622
[step: 501 - Thu Oct 26 06:39:11 2023] - loss:0.00929985
[step: 601 - Thu Oct 26 06:39:17 2023] - loss:0.00813591
