Chapter 44
文本分类实例
Notebooktransformers25 cells
文本分类实例
Step1 导入相关包
In [ ]python · cell 3
python
from transformers import AutoTokenizer, AutoModelForSequenceClassification, Trainer, TrainingArguments, BertTokenizer, BertForSequenceClassification
from datasets import load_datasetStep2 加载数据集
In [ ]python · cell 5
python
dataset = load_dataset("csv", data_files="./ChnSentiCorp_htl_all.csv", split="train")
dataset = dataset.filter(lambda x: x["review"] is not None)
datasetStep3 划分数据集
In [ ]python · cell 7
python
datasets = dataset.train_test_split(test_size=0.1)
datasetsStep4 数据集预处理
In [ ]python · cell 9
python
import torch
tokenizer = BertTokenizer.from_pretrained("hfl/rbt3")
def process_function(examples):
tokenized_examples = tokenizer(examples["review"], max_length=128, truncation=True)
tokenized_examples["labels"] = examples["label"]
return tokenized_examples
tokenized_datasets = datasets.map(process_function, batched=True, remove_columns=datasets["train"].column_names)
tokenized_datasetsStep5 创建模型
In [ ]python · cell 11
python
model = BertForSequenceClassification.from_pretrained("hfl/rbt3")In [ ]python · cell 12
python
model.configStep6 创建评估函数
In [ ]python · cell 14
python
import evaluate
acc_metric = evaluate.load("./metric_accuracy.py")
f1_metirc = evaluate.load("./metric_f1.py")In [ ]python · cell 15
python
def eval_metric(eval_predict):
predictions, labels = eval_predict
predictions = predictions.argmax(axis=-1)
acc = acc_metric.compute(predictions=predictions, references=labels)
f1 = f1_metirc.compute(predictions=predictions, references=labels)
acc.update(f1)
return accStep7 创建TrainingArguments
In [ ]python · cell 17
python
train_args = TrainingArguments(output_dir="./checkpoints", # 输出文件夹
per_device_train_batch_size=64, # 训练时的batch_size
per_device_eval_batch_size=128, # 验证时的batch_size
logging_steps=10, # log 打印的频率
evaluation_strategy="epoch", # 评估策略
save_strategy="epoch", # 保存策略
save_total_limit=3, # 最大保存数
learning_rate=2e-5, # 学习率
weight_decay=0.01, # weight_decay
metric_for_best_model="f1", # 设定评估指标
load_best_model_at_end=True) # 训练完成后加载最优模型In [ ]python · cell 18
python
hasattr(train_args, "_n_gpu")
train_args.__dict__Step8 创建Trainer
In [ ]python · cell 20
python
from transformers import DataCollatorWithPadding
trainer = Trainer(model=model,
args=train_args,
train_dataset=tokenized_datasets["train"],
eval_dataset=tokenized_datasets["test"],
data_collator=DataCollatorWithPadding(tokenizer=tokenizer),
compute_metrics=eval_metric)Step9 模型训练
In [ ]python · cell 22
python
trainer.train()Step10 模型评估
In [ ]python · cell 24
python
trainer.evaluate(tokenized_datasets["test"])In [ ]python · cell 25
python