【深度学习前沿应用】文本分类Fine-Tunning
前言
应用BERT模型做短文本情绪分类
#导入相关的模块
import paddle
import paddlenlp as ppnlp
from paddlenlp.data import Stack, Pad, Tuple
import paddle.nn.functional as F
import numpy as np
from functools import partial #partial()函数可以用来固定某些参数值,并返回一个新的callable对象
ppnlp.__version__
一、数据加载及预处理
(一)、数据导入
数据集为公开中文情感分析数据集ChnSenticorp。使用PaddleNLP的.datasets.ChnSentiCorp.get_datasets方法即可以加载该数据集。
#采用paddlenlp内置的ChnSentiCorp语料,该语料主要可以用来做情感分类。训练集用来训练模型,验证集用来选择模型,测试集用来评估模型泛化性能。
train_ds, dev_ds, test_ds = ppnlp.datasets.ChnSentiCorp.get_datasets(['train','dev','test'])
#获得标签列表
label_list = train_ds.get_labels()
#看看数据长什么样子,分别打印训练集、验证集、测试集的前3条数据。
print("训练集数据:{}\n".format(train_ds[0:1]))
print("验证集数据:{}\n".format(dev_ds[0:1]))
print("测试集数据:{}\n".format(test_ds[0:1]))
print("训练集样本个数:{}".format(len(train_ds)))
print("验证集样本个数:{}".format(len(dev_ds)))
print("测试集样本个数:{}".format(len(test_ds)))
输出结果如下图1所示:
(二)、数据预处理
#调用ppnlp.transformers.BertTokenizer进行数据处理,tokenizer可以把原始输入文本转化成模型model可接受的输入数据格式。
tokenizer = ppnlp.transformers.BertTokenizer.from_pretrained("bert-base-chinese")
#数据预处理
def convert_example(example,tokenizer,label_list,max_seq_length=256,is_test=False):
if is_test:
text = example
else:
text, label = example
#tokenizer.encode方法能够完成切分token,映射token ID以及拼接特殊token
encoded_inputs = tokenizer.encode(text=text, max_seq_len=max_seq_length)
# print('===================')
# print(encoded_inputs)
input_ids = encoded_inputs["input_ids"]
segment_ids = encoded_inputs["token_type_ids"]
if not is_test:
label_map = {}
for (i, l) in enumerate(label_list):
label_map[l] = i
label = label_map[label]
label = np.array([label], dtype="int64")
return input_ids, segment_ids, label
else:
return input_ids, segment_ids
#数据迭代器构造方法
def create_dataloader(dataset, trans_fn=None, mode='train', batch_size=1, use_gpu=False, pad_token_id=0, batchify_fn=None):
if trans_fn:
dataset = dataset.apply(trans_fn, lazy=True)
if mode == 'train' and use_gpu:
sampler = paddle.io.DistributedBatchSampler(dataset=dataset, batch_size=batch_size, shuffle=True)
else:
shuffle = True if mode == 'train' else False #如果不是训练集,则不打乱顺序
sampler = paddle.io.BatchSampler(dataset=dataset, batch_size=batch_size, shuffle=shuffle) #生成一个取样器
dataloader = paddle.io.DataLoader(dataset, batch_sampler=sampler, return_list=True, collate_fn=batchify_fn)
return dataloader
#使用partial()来固定convert_example函数的tokenizer, label_list, max_seq_length, is_test等参数值
trans_fn = partial(convert_example, tokenizer=tokenizer, label_list=label_list, max_seq_length=128, is_test=False)
batchify_fn = lambda samples, fn=Tuple(Pad(axis=0,pad_val=tokenizer.pad_token_id), Pad(axis=0, pad_val=tokenizer.pad_token_id), Stack(dtype="int64")):[data for data in fn(samples)]
#训练集迭代器
train_loader = create_dataloader(train_ds, mode='train', batch_size=64, batchify_fn=batchify_fn, trans_fn=trans_fn)
#验证集迭代器
dev_loader = create_dataloader(dev_ds, mode='dev', batch_size=64, batchify_fn=batchify_fn, trans_fn=trans_fn)
#测试集迭代器
test_loader = create_dataloader(test_ds, mode='test', batch_size=64, batchify_fn=batchify_fn, trans_fn=trans_fn)
二、BERT预训练模型加载
#加载预训练模型Bert用于文本分类任务的Fine-tune网络BertForSequenceClassification, 它在BERT模型后接了一个全连接层进行分类。
#由于本任务中的情感分类是二分类问题,设定num_classes为2
model = ppnlp.transformers.BertForSequenceClassification.from_pretrained("bert-base-chinese", num_classes=2)
三、训练模型
(一)、设置训练超参数
#设置训练超参数
#学习率
learning_rate = 1e-5
#训练轮次
epochs = 8
#学习率预热比率
warmup_proption = 0.1
#权重衰减系数
weight_decay = 0.01
num_training_steps = len(train_loader) * epochs
num_warmup_steps = int(warmup_proption * num_training_steps)
def get_lr_factor(current_step):
if current_step < num_warmup_steps:
return float(current_step) / float(max(1, num_warmup_steps))
else:
return max(0.0,
float(num_training_steps - current_step) /
float(max(1, num_training_steps - num_warmup_steps)))
#学习率调度器
lr_scheduler = paddle.optimizer.lr.LambdaDecay(learning_rate, lr_lambda=lambda current_step: get_lr_factor(current_step))
#优化器
optimizer = paddle.optimizer.AdamW(
learning_rate=lr_scheduler,
parameters=model.parameters(),
weight_decay=weight_decay,
apply_decay_param_fun=lambda x: x in [
p.name for n, p in model.named_parameters()
if not any(nd in n for nd in ["bias", "norm"])
])
#损失函数
criterion = paddle.nn.loss.CrossEntropyLoss()
#评估函数
metric = paddle.metric.Accuracy()
(二)、评估函数
#评估函数
def evaluate(model, criterion, metric, data_loader):
model.eval()
metric.reset()
losses = []
for batch in data_loader:
input_ids, segment_ids, labels = batch
logits = model(input_ids, segment_ids)
loss = criterion(logits, labels)
losses.append(loss.numpy())
correct = metric.compute(logits, labels)
metric.update(correct)
accu = metric.accumulate()
print("eval loss: %.5f, accu: %.5f" % (np.mean(losses), accu))
model.train()
metric.reset()
(三)、模型训练
#开始训练
global_step = 0
for epoch in range(1, epochs + 1):
for step, batch in enumerate(train_loader): #从训练数据迭代器中取数据
# print(batch)
input_ids, segment_ids, labels = batch
logits = model(input_ids, segment_ids)
loss = criterion(logits, labels) #计算损失
probs = F.softmax(logits, axis=1)
correct = metric.compute(probs, labels)
metric.update(correct)
acc = metric.accumulate()
global_step += 1
if global_step % 50 == 0 :
print("global step %d, epoch: %d, batch: %d, loss: %.5f, acc: %.5f" % (global_step, epoch, step, loss, acc))
loss.backward()
optimizer.step()
lr_scheduler.step()
optimizer.clear_gradients()
evaluate(model, criterion, metric, dev_loader)
四、模型预测
def predict(model, data, tokenizer, label_map, batch_size=1):
examples = []
for text in data:
input_ids, segment_ids = convert_example(text, tokenizer, label_list=label_map.values(), max_seq_length=128, is_test=True)
examples.append((input_ids, segment_ids))
batchify_fn = lambda samples, fn=Tuple(Pad(axis=0, pad_val=tokenizer.pad_token_id), Pad(axis=0, pad_val=tokenizer.pad_token_id)): fn(samples)
batches = []
one_batch = []
for example in examples:
one_batch.append(example)
if len(one_batch) == batch_size:
batches.append(one_batch)
one_batch = []
if one_batch:
batches.append(one_batch)
results = []
model.eval()
for batch in batches:
input_ids, segment_ids = batchify_fn(batch)
input_ids = paddle.to_tensor(input_ids)
segment_ids = paddle.to_tensor(segment_ids)
logits = model(input_ids, segment_ids)
probs = F.softmax(logits, axis=1)
idx = paddle.argmax(probs, axis=1).numpy()
idx = idx.tolist()
labels = [label_map[i] for i in idx]
results.extend(labels)
return results
data = ['这个商品虽然看着样式挺好看的,但是不耐用。', '这个老师讲课水平挺高的。']
label_map = {0: '负向情绪', 1: '正向情绪'}
predictions = predict(model, data, tokenizer, label_map, batch_size=32)
for idx, text in enumerate(data):
print('预测文本: {} \n情绪标签: {}'.format(text, predictions[idx]))
输出结果如下图2所示:
总结
本系列文章内容为根据清华社出版的《机器学习实践》所作的相关笔记和感悟,其中代码均为基于百度飞桨开发,若有任何侵权和不妥之处,请私信于我,定积极配合处理,看到必回!!!
最后,引用本次活动的一句话,来作为文章的结语~( ̄▽ ̄~)~:
【学习的最大理由是想摆脱平庸,早一天就多一份人生的精彩;迟一天就多一天平庸的困扰。】