100字范文 > 新闻主题分类任务——torchtext 库进行文本分类

新闻主题分类任务——torchtext 库进行文本分类

时间：2020-08-17 15:14:07

简介

使用浅层网络构建新闻主题分类器。

以一段新闻报道中的文本描述内容为输入, 使用模型帮助我们判断它最有可能属于哪一种类型的新闻, 这是典型的文本分类问题, 我们这里假定每种类型是互斥的, 即文本描述有且只有一种类型。

导入相关的torch工具包

import timeimport torchimport torch.nn as nnfrom torchtext.datasets import AG_NEWSfrom torchtext.data.utils import get_tokenizerfrom torchtext.vocab import build_vocab_from_iteratorfrom torch.utils.data import DataLoaderfrom torch.utils.data.dataset import random_splitfrom torchtext.data.functional import to_map_style_datasetfrom TextClassificationModule import TextClassificationModule

访问原始数据集迭代器

torchtext 库提供了一些原始数据集迭代器，它们产生原始文本字符串。例如，AG_NEWS数据集迭代器将原始数据生成为标签和文本的元组。

# 可用设备检测, 有GPU的话将优先使用GPUdevice = torch.device("cuda" if torch.cuda.is_available() else "cpu")# 基本的英文分词器tokenizer = get_tokenizer('basic_english')# 训练数据加载器train_iter = AG_NEWS(split="train")test_iter = AG_NEWS(split="test")

对读取到的数据进行测试，该读取的数据是从网上自动下载到缓存，其中读取到的 train_iter 和 test_iter 为训练集和测试集，且均为迭代器类型。

print('test:')train_data = iter(train_iter)test_data = iter(test_iter)print(next(train_data))print(next(test_data))

运行结果

test:(3, "Wall St. Bears Claw Back Into the Black (Reuters) Reuters - Short-sellers, Wall Street's dwindling\\band of ultra-cynics, are seeing green again.")(3, "Fears for T N pension after talks Unions representing workers at Turner Newall say they are 'disappointed' after talks with stricken parent firm Federal Mogul.")

使用原始训练数据集构建词汇表

其中分词生成器中的 “_” 表示一个不用的变量即类别，text 表示新闻文本，如：

_ = 3

text = Wall St. Bears Claw Back Into the Black (Reuters) Reuters - Short-sellers, Wall Street's dwindling\\band of ultra-cynics, are seeing green again.

python 中 yield 的作用就是把一个函数变成一个 generator，带有 yield 的函数不再是一个普通函数，Python 解释器会将其视为一个 generator，调用 fab(5) 不会执行 fab 函数，而是返回一个 iterable 对象。

示例

def yield_test(n): for i in range(n): yield call(i) print("i=",i) #做一些其它的事情print("do something.")print("end.") def call(i): return i*2 #使用for循环 for i in yield_test(5): print(i,",")

运行结果

0 , i= 0 2 , i= 1 4 , i= 2 6 , i= 3 8 , i= 4 do something. end.

使用原始训练数据集构建词汇表

# 分词生成器def yield_tokens(data_iter):for _, text in data_iter:yield tokenizer(text)# 根据训练数据构建词汇表，torchtext.vocabvocab = build_vocab_from_iterator(yield_tokens(train_iter), specials=["<unk>"])# 设置默认索引，当某个单词不在词汇表 vocab 时（OOV)，返回该单词索引vocab.set_default_index(vocab["<unk>"])# 词汇表会将 token 映射到词汇表中的索引上print(vocab(["here", "is", "an", "example"]))# 构建数据加载器 dataloader# text_pipeline 将一个文本字符串转换为整数 List, List 中每项对应词汇表 vocab 中的单词的索引号text_pipeline = lambda x: vocab(tokenizer(x))# label_pipeline 将 label 转换为整数label_pipeline = lambda x: int(x) - 1# pipeline exampleprint(text_pipeline("hello world! I'am happy"))print(label_pipeline("10"))

运行结果

[475, 21, 30, 5297][12544, 50, 764, 282, 16, 1913, 2734]9

生成数据批处理和迭代器

def collate_batch(batch):label_list, text_list, offsets = [], [], [0]for (_label, _text) in batch:label_list.append(label_pipeline(_label))processed_text = torch.tensor(text_pipeline(_text), dtype=torch.int64)text_list.append(processed_text)offsets.append(processed_text.size(0))label_list = torch.tensor(label_list, dtype=torch.int64)offsets = torch.tensor(offsets[:-1]).cumsum(dim=0)text_list = torch.cat(text_list)return label_list.to(device), text_list.to(device), offsets.to(device)# 加载数据集合，转换为张量dataloader = DataLoader(train_iter, batch_size=8, shuffle=False, collate_fn=collate_batch)

定义模型

该模型由 nn.EmbeddingBag 层和用于分类目的的线性层组成。nn.EmbeddingBag 使用默认模式“mean”计算嵌入“bag”的平均值。尽管此处的文本条目具有不同的长度，但 nn.EmbeddingBag 模块在此处不需要填充，因为文本长度保存在偏移量中。

nn.EmbeddingBag 可以提高性能和内存效率以处理一系列张量。

import torch.nn as nnclass TextClassificationModule(nn.Module):def __init__(self, vocab_size, embed_dim, num_class):"""文本分类模型description: 类的初始化函数:param vocab_size: 整个语料包含的不同词汇总数:param embed_dim: 指定词嵌入的维度:param num_class: 文本分类的类别总数"""super(TextClassificationModule, self).__init__()# 实例化embedding层, sparse=True代表每次对该层求解梯度时, 只更新部分权重self.embedding = nn.EmbeddingBag(vocab_size, embed_dim, sparse=True)# 实例化全连接层, 参数分别是embed_dim和num_classself.fc = nn.Linear(embed_dim, num_class)# 为各层初始化权重self.init_weights()def init_weights(self):"""初始化权重函数"""# 指定初始权重的取值范围数initrange = 0.5# 各层的权重参数都是初始化为均匀分布self.embedding.weight.data.uniform_(-initrange, initrange)self.fc.weight.data.uniform_(-initrange, initrange)# 偏置初始化为0self.fc.bias.data.zero_()def forward(self, text, offsets):""":param text: 文本数值映射后的结果:return: 与类别数尺寸相同的张量, 用以判断文本类别"""embedded = self.embedding(text, offsets)return self.fc(embedded)

定义函数来训练模型和评估结果

def train(dataloader):model.train()total_acc, total_count = 0, 0log_interval = 500start_time = time.time()for idx, (label, text, offsets) in enumerate(dataloader):optimizer.zero_grad()predicted_label = model(text, offsets)loss = criterion(predicted_label, label)loss.backward()torch.nn.utils.clip_grad_norm_(model.parameters(), 0.1)optimizer.step()total_acc += (predicted_label.argmax(1) == label).sum().item()total_count += label.size(0)if idx % log_interval == 0 and idx > 0:elapsed = time.time() - start_timeprint('| epoch {:3d} | {:5d}/{:5d} batches ''| accuracy {:8.3f}'.format(epoch, idx, len(dataloader),total_acc / total_count))total_acc, total_count = 0, 0start_time = time.time()def evaluate(dataloader):model.eval()total_acc, total_count = 0, 0with torch.no_grad():for idx, (label, text, offsets) in enumerate(dataloader):predicted_label = model(text, offsets)loss = criterion(predicted_label, label)total_acc += (predicted_label.argmax(1) == label).sum().item()total_count += label.size(0)return total_acc / total_count

实例化并运行模型

# 加载数据集合，转换为张量dataloader = DataLoader(train_iter, batch_size=8, shuffle=False, collate_fn=collate_batch)# 一个嵌入维度为 64 的模型。词汇大小等于词汇实例的长度。类的数量等于标签的数量，num_class = len(set([label for (label, text) in train_iter]))vocab_size = len(vocab)emsize = 64model = TextClassificationModule(vocab_size, emsize, num_class).to(device)# 训练轮数EPOCHS = 10# 学习率LR = 5# 训练数据规模BATCH_SIZE = 64# 交叉熵损失函数criterion = torch.nn.CrossEntropyLoss()# 优化器optimizer = torch.optim.SGD(model.parameters(), lr=LR)# 调整学习率机制scheduler = torch.optim.lr_scheduler.StepLR(optimizer, 1, gamma=0.1)total_accu = Nonetrain_dataset = to_map_style_dataset(train_iter)test_dataset = to_map_style_dataset(test_iter)# 划分训练集中5%的数据最为验证集num_train = int(len(train_dataset) * 0.95)split_train_, split_valid_ = random_split(train_dataset, [num_train, len(train_dataset) - num_train])train_dataloader = DataLoader(split_train_, batch_size=BATCH_SIZE, shuffle=True, collate_fn=collate_batch)valid_dataloader = DataLoader(split_valid_, batch_size=BATCH_SIZE, shuffle=True, collate_fn=collate_batch)test_dataloader = DataLoader(test_dataset, batch_size=BATCH_SIZE, shuffle=True, collate_fn=collate_batch)for epoch in range(1, EPOCHS + 1):epoch_start_time = time.time()train(train_dataloader)accu_val = evaluate(valid_dataloader)if total_accu is not None and total_accu > accu_val:scheduler.step()else:total_accu = accu_valprint('-' * 59)print('| end of epoch {:3d} | time: {:5.2f}s | ''valid accuracy {:8.3f} '.format(epoch,time.time() - epoch_start_time,accu_val))print('-' * 59)

运行结果

| epoch 1 | 500/ 1782 batches | accuracy 0.689| epoch 1 | 1000/ 1782 batches | accuracy 0.856| epoch 1 | 1500/ 1782 batches | accuracy 0.873-----------------------------------------------------------| end of epoch 1 | time: 23.38s | valid accuracy 0.879 -----------------------------------------------------------| epoch 2 | 500/ 1782 batches | accuracy 0.896| epoch 2 | 1000/ 1782 batches | accuracy 0.904| epoch 2 | 1500/ 1782 batches | accuracy 0.900-----------------------------------------------------------| end of epoch 2 | time: 32.21s | valid accuracy 0.891 -----------------------------------------------------------| epoch 3 | 500/ 1782 batches | accuracy 0.915| epoch 3 | 1000/ 1782 batches | accuracy 0.916| epoch 3 | 1500/ 1782 batches | accuracy 0.915-----------------------------------------------------------| end of epoch 3 | time: 36.85s | valid accuracy 0.899 -----------------------------------------------------------| epoch 4 | 500/ 1782 batches | accuracy 0.925| epoch 4 | 1000/ 1782 batches | accuracy 0.925| epoch 4 | 1500/ 1782 batches | accuracy 0.922-----------------------------------------------------------| end of epoch 4 | time: 20.15s | valid accuracy 0.897 -----------------------------------------------------------| epoch 5 | 500/ 1782 batches | accuracy 0.937| epoch 5 | 1000/ 1782 batches | accuracy 0.938| epoch 5 | 1500/ 1782 batches | accuracy 0.936-----------------------------------------------------------| end of epoch 5 | time: 28.52s | valid accuracy 0.905 -----------------------------------------------------------| epoch 6 | 500/ 1782 batches | accuracy 0.939| epoch 6 | 1000/ 1782 batches | accuracy 0.938| epoch 6 | 1500/ 1782 batches | accuracy 0.941-----------------------------------------------------------| end of epoch 6 | time: 33.47s | valid accuracy 0.905 -----------------------------------------------------------| epoch 7 | 500/ 1782 batches | accuracy 0.940| epoch 7 | 1000/ 1782 batches | accuracy 0.941| epoch 7 | 1500/ 1782 batches | accuracy 0.939-----------------------------------------------------------| end of epoch 7 | time: 20.75s | valid accuracy 0.904 -----------------------------------------------------------| epoch 8 | 500/ 1782 batches | accuracy 0.941| epoch 8 | 1000/ 1782 batches | accuracy 0.941| epoch 8 | 1500/ 1782 batches | accuracy 0.940-----------------------------------------------------------| end of epoch 8 | time: 27.11s | valid accuracy 0.906 -----------------------------------------------------------| epoch 9 | 500/ 1782 batches | accuracy 0.942| epoch 9 | 1000/ 1782 batches | accuracy 0.942| epoch 9 | 1500/ 1782 batches | accuracy 0.942-----------------------------------------------------------| end of epoch 9 | time: 34.83s | valid accuracy 0.906 -----------------------------------------------------------| epoch 10 | 500/ 1782 batches | accuracy 0.942| epoch 10 | 1000/ 1782 batches | accuracy 0.942| epoch 10 | 1500/ 1782 batches | accuracy 0.940-----------------------------------------------------------| end of epoch 10 | time: 22.78s | valid accuracy 0.906 -----------------------------------------------------------

使用测试数据集评估模型

print('Checking the results of test dataset.')accu_test = evaluate(test_dataloader)print('test accuracy {:8.3f}'.format(accu_test))

运行结果

Checking the results of test dataset.test accuracy 0.906

测试随机新闻

# 测试随机新闻# 使用迄今为止最好的模型并测试高尔夫新闻。ag_news_label = {1: "World",2: "Sports",3: "Business",4: "Sci/Tec"}def predict(text, text_pipeline):with torch.no_grad():text = torch.tensor(text_pipeline(text))output = model(text, torch.tensor([0]))return output.argmax(1).item() + 1ex_text_str = "MEMPHIS, Tenn. – Four days ago, Jon Rahm was \enduring the season’s worst weather conditions on Sunday at The \Open on his way to a closing 75 at Royal Portrush, which \considering the wind and the rain was a respectable showing. \Thursday’s first round at the WGC-FedEx St. Jude Invitational \was another story. With temperatures in the mid-80s and hardly any \wind, the Spaniard was 13 strokes better in a flawless round. \Thanks to his best putting performance on the PGA Tour, Rahm \finished with an 8-under 62 for a three-stroke lead, which \was even more impressive considering he’d never played the \front nine at TPC Southwind."model = model.to("cpu")print("This is a %s news" % ag_news_label[predict(ex_text_str, text_pipeline)])

运行结果

This is a Sports news

完整代码

import timeimport torchimport torch.nn as nnfrom torchtext.datasets import AG_NEWSfrom torchtext.data.utils import get_tokenizerfrom torchtext.vocab import build_vocab_from_iteratorfrom torch.utils.data import DataLoaderfrom torch.utils.data.dataset import random_splitfrom torchtext.data.functional import to_map_style_dataset# 可用设备检测, 有GPU的话将优先使用GPUdevice = torch.device("cuda" if torch.cuda.is_available() else "cpu")# 基本的英文分词器tokenizer = get_tokenizer('basic_english')# 训练数据加载器train_iter = AG_NEWS(split="train")test_iter = AG_NEWS(split="test")# print('test:')# train_data = iter(train_iter)# test_data = iter(test_iter)# print(next(train_data))# print(next(test_data))# 分词生成器def yield_tokens(data_iter):for _, text in data_iter:yield tokenizer(text)# 根据训练数据构建词汇表vocab = build_vocab_from_iterator(yield_tokens(train_iter), specials=["<unk>"])# 设置默认索引，当某个单词不在词汇表 vocab 时（OOV)，返回该单词索引vocab.set_default_index(vocab["<unk>"])# 词汇表会将 token 映射到词汇表中的索引上# print(vocab(["here", "is", "an", "example"]))# 构建数据加载器 dataloader# text_pipeline 将一个文本字符串转换为整数 List, List 中每项对应词汇表 vocab 中的单词的索引号text_pipeline = lambda x: vocab(tokenizer(x))# label_pipeline 将 label 转换为整数label_pipeline = lambda x: int(x) - 1# pipeline example# print(text_pipeline("hello world! I'am happy"))# print(label_pipeline("10"))# 模型class TextClassificationModule(nn.Module):def __init__(self, vocab_size, embed_dim, num_class):"""文本分类模型description: 类的初始化函数:param vocab_size: 整个语料包含的不同词汇总数:param embed_dim: 指定词嵌入的维度:param num_class: 文本分类的类别总数"""super(TextClassificationModule, self).__init__()# 实例化embedding层, sparse=True代表每次对该层求解梯度时, 只更新部分权重self.embedding = nn.EmbeddingBag(vocab_size, embed_dim, sparse=True)# 实例化全连接层, 参数分别是embed_dim和num_classself.fc = nn.Linear(embed_dim, num_class)# 为各层初始化权重self.init_weights()def init_weights(self):"""初始化权重函数"""# 指定初始权重的取值范围数initrange = 0.5# 各层的权重参数都是初始化为均匀分布self.embedding.weight.data.uniform_(-initrange, initrange)self.fc.weight.data.uniform_(-initrange, initrange)# 偏置初始化为0self.fc.bias.data.zero_()def forward(self, text, offsets):""":param text: 文本数值映射后的结果:return: 与类别数尺寸相同的张量, 用以判断文本类别"""embedded = self.embedding(text, offsets)return self.fc(embedded)def collate_batch(batch):label_list, text_list, offsets = [], [], [0]for (_label, _text) in batch:label_list.append(label_pipeline(_label))processed_text = torch.tensor(text_pipeline(_text), dtype=torch.int64)text_list.append(processed_text)offsets.append(processed_text.size(0))label_list = torch.tensor(label_list, dtype=torch.int64)offsets = torch.tensor(offsets[:-1]).cumsum(dim=0)text_list = torch.cat(text_list)return label_list.to(device), text_list.to(device), offsets.to(device)def train(dataloader):model.train()total_acc, total_count = 0, 0log_interval = 500start_time = time.time()for idx, (label, text, offsets) in enumerate(dataloader):optimizer.zero_grad()predicted_label = model(text, offsets)loss = criterion(predicted_label, label)loss.backward()torch.nn.utils.clip_grad_norm_(model.parameters(), 0.1)optimizer.step()total_acc += (predicted_label.argmax(1) == label).sum().item()total_count += label.size(0)if idx % log_interval == 0 and idx > 0:elapsed = time.time() - start_timeprint('| epoch {:3d} | {:5d}/{:5d} batches ''| accuracy {:8.3f}'.format(epoch, idx, len(dataloader),total_acc / total_count))total_acc, total_count = 0, 0start_time = time.time()def evaluate(dataloader):model.eval()total_acc, total_count = 0, 0with torch.no_grad():for idx, (label, text, offsets) in enumerate(dataloader):predicted_label = model(text, offsets)loss = criterion(predicted_label, label)total_acc += (predicted_label.argmax(1) == label).sum().item()total_count += label.size(0)return total_acc / total_count# 加载数据集合，转换为张量dataloader = DataLoader(train_iter, batch_size=8, shuffle=False, collate_fn=collate_batch)# 一个嵌入维度为 64 的模型。词汇大小等于词汇实例的长度。类的数量等于标签的数量，num_class = len(set([label for (label, text) in train_iter]))vocab_size = len(vocab)emsize = 64model = TextClassificationModule(vocab_size, emsize, num_class).to(device)# 训练轮数EPOCHS = 10# 学习率LR = 5# 训练数据规模BATCH_SIZE = 64# 交叉熵损失函数criterion = torch.nn.CrossEntropyLoss()# 优化器optimizer = torch.optim.SGD(model.parameters(), lr=LR)# 调整学习率机制scheduler = torch.optim.lr_scheduler.StepLR(optimizer, 1, gamma=0.1)total_accu = Nonetrain_dataset = to_map_style_dataset(train_iter)test_dataset = to_map_style_dataset(test_iter)# 划分训练集中5%的数据最为验证集num_train = int(len(train_dataset) * 0.95)split_train_, split_valid_ = random_split(train_dataset, [num_train, len(train_dataset) - num_train])train_dataloader = DataLoader(split_train_, batch_size=BATCH_SIZE, shuffle=True, collate_fn=collate_batch)valid_dataloader = DataLoader(split_valid_, batch_size=BATCH_SIZE, shuffle=True, collate_fn=collate_batch)test_dataloader = DataLoader(test_dataset, batch_size=BATCH_SIZE, shuffle=True, collate_fn=collate_batch)for epoch in range(1, EPOCHS + 1):epoch_start_time = time.time()train(train_dataloader)accu_val = evaluate(valid_dataloader)if total_accu is not None and total_accu > accu_val:scheduler.step()else:total_accu = accu_valprint('-' * 59)print('| end of epoch {:3d} | time: {:5.2f}s | ''valid accuracy {:8.3f} '.format(epoch,time.time() - epoch_start_time,accu_val))print('-' * 59)'''使用测试数据集评估模型'''print('Checking the results of test dataset.')accu_test = evaluate(test_dataloader)print('test accuracy {:8.3f}'.format(accu_test))# 测试随机新闻# 使用迄今为止最好的模型并测试高尔夫新闻。ag_news_label = {1: "World",2: "Sports",3: "Business",4: "Sci/Tec"}def predict(text, text_pipeline):with torch.no_grad():text = torch.tensor(text_pipeline(text))output = model(text, torch.tensor([0]))return output.argmax(1).item() + 1ex_text_str = "MEMPHIS, Tenn. – Four days ago, Jon Rahm was \enduring the season’s worst weather conditions on Sunday at The \Open on his way to a closing 75 at Royal Portrush, which \considering the wind and the rain was a respectable showing. \Thursday’s first round at the WGC-FedEx St. Jude Invitational \was another story. With temperatures in the mid-80s and hardly any \wind, the Spaniard was 13 strokes better in a flawless round. \Thanks to his best putting performance on the PGA Tour, Rahm \finished with an 8-under 62 for a three-stroke lead, which \was even more impressive considering he’d never played the \front nine at TPC Southwind."model = model.to("cpu")print("This is a %s news" % ag_news_label[predict(ex_text_str, text_pipeline)])

运行结果