利用Torchtext 0.12.0以上版本(实际0.14.0)进行IMDB评论情感分类(消极或积极)

随手笔记前言

利用Torchtext 0.12.0以上版本(实际0.14.0)进行IMDB评论情感分类(消极或积极)

---

一、写该随手目的？

由于Torchtext0.12.0版本大更新，现使用Torchtext0.12.0以上版本(实际0.14.0)进行IMDB评论态度分类。
由于在使用以下新的torchtext进行IMDB数据下载时，常常因为网络问题不能下载，现附上IMDB数据集。
链接：https://pan.baidu.com/s/1-2a7QwTzcRonO3pG3DZXbw
提取码：yyds

from torchtext.datasets import IMDB
train_iter, test_iter = IMDB()

数据集下载后，放在 C:\Users\<你自己的用户名>\.cache\torch\text\datasets\IMDB\
目录下即可。

二、随手笔记内容

1.代码

代码如下：

import torch
from torchtext.datasets import IMDB
import torch.nn as nn

from torchtext.data.utils import get_tokenizer
from torchtext.vocab import build_vocab_from_iterator

tokenizer = get_tokenizer('basic_english')  # 分割器
train_iter = IMDB(split='train')  # 下载IMDB数据集训练集


def yield_tokens(data_iter):
    for _, text in data_iter:
        yield tokenizer(text)


vocab = build_vocab_from_iterator(yield_tokens(train_iter), specials=['<unk>'])  # 建立词表
vocab.set_default_index(vocab['<unk>'])

# print(vocab(['here', 'is', 'an', 'example']))

text_pipeline = lambda x: vocab(tokenizer(x))  # 处理文本的管道
label_pipeline = lambda x: int(x) - 1  # 处理label的管道


from torch.utils.data import DataLoader

device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")


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)


dataloader = DataLoader(train_iter, batch_size=8, shuffle=False, collate_fn=collate_batch)  # 形成批处理的dataloader


class TextClassificationModel(nn.Module):  # 分类模型

    def __init__(self, vocab_size, embed_dim, num_class):
        super(TextClassificationModel, self).__init__()
        self.embedding = nn.EmbeddingBag(vocab_size, embed_dim, sparse=False)
        self.fc = nn.Linear(embed_dim, num_class)
        self.init_weights()

    def init_weights(self):
        initrange = 0.5
        self.embedding.weight.data.uniform_(-initrange, initrange)
        self.fc.weight.data.uniform_(-initrange, initrange)
        self.fc.bias.data.zero_()

    def forward(self, text, offsets):
        embedded = self.embedding(text, offsets)
        return self.fc(embedded)


num_class = len(set([label for (label, text) in train_iter]))
vocab_size = len(vocab)
emsize = 64
model = TextClassificationModel(vocab_size, emsize, num_class).to(device)

import time


def train(dataloader, optimizer, criterion, epoch):  # 训练函数
    model.train()
    total_acc, total_count = 0, 0
    log_interval = 50
    start_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_time
            print('| epoch {:3d} | {:5d}/{:5d} batches '
                  '| accuracy {:8.3f}'.format(epoch, idx, len(dataloader),
                                              total_acc / total_count))
            total_acc, total_count = 0, 0
            start_time = time.time()


def evaluate(dataloader):  # 评价函数
    model.eval()
    total_acc, total_count = 0, 0

    with torch.no_grad():
        for idx, (label, text, offsets) in enumerate(dataloader):
            predicted_label = model(text, offsets)
            total_acc += (predicted_label.argmax(1) == label).sum().item()
            total_count += label.size(0)
    return total_acc / total_count


from torch.utils.data.dataset import random_split
from torchtext.data.functional import to_map_style_dataset

# 超参
EPOCHS = 10  # epoch
LR = 5  # 学习率
BATCH_SIZE = 64  # batch大小

criterion = torch.nn.CrossEntropyLoss()  # 这里也可以二分类损失
optimizer = torch.optim.SGD(model.parameters(), lr=LR)  # 这里也可以定义Adam优化
scheduler = torch.optim.lr_scheduler.StepLR(optimizer, 1.0, gamma=0.1)
total_accu = None

train_iter, test_iter = IMDB()  # IMDB训练集测试集都下载
train_dataset = to_map_style_dataset(train_iter)
test_dataset = to_map_style_dataset(test_iter)
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)
# 训练
for epoch in range(1, EPOCHS + 1):
    epoch_start_time = time.time()
    train(train_dataloader, optimizer, criterion, epoch)
    accu_val = evaluate(valid_dataloader)
    if total_accu is not None and total_accu > accu_val:
        scheduler.step()
    else:
        total_accu = accu_val
    print('-' * 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)

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


# 实际演示
ag_news_label = {1: "neg",
                 2: "pos"}


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() + 1


ex_text_str = "This movie is rubbish, I don't like it very much."

model = model.to("cpu")
print()
print('-----------------------------------------------')
print("This is a %s review" % ag_news_label[predict(ex_text_str, text_pipeline)])

2.展示结果

以上只是经过了10个epoch的训练，同时没有做太多的优化，毕竟这个只是做一个Demo，后续可以在各个方面做优化，来提高准确率。

---

总结

以上即为：利用Torchtext 0.12.0以上版本(实际0.14.0)进行IMDB评论情感分类(消极或积极)