Chinese-Text-Classification

KO | EN

gitlite — search

#javascript #python #hacktoberfest #react #ai #typescript #llm #go #golang #android #machine-learning #rust #deep-learning #linux

Chinese-Text-Classification

★ 9 Open GitHub ↗

新闻标题文本分类

// repository documentation

Was this content helpful?

★ 0 (0 ratings)

![](https://img.shields.io/badge/license-MIT-blue.svg) 
![](https://img.shields.io/badge/Python-3.6.12-blue.svg)
![](https://img.shields.io/badge/torch-1.7.0-brightgreen.svg)
![](https://img.shields.io/badge/transformers-4.4.1-brightgreen.svg)
![](https://img.shields.io/badge/scikitlearn-0.24.0-brightgreen.svg)
![](https://img.shields.io/badge/tqdm-4.49.0-brightgreen.svg)
![](https://img.shields.io/badge/jsonlines-2.0.0-brightgreen.svg)
![](https://img.shields.io/badge/loguru-0.5.3-brightgreen.svg)

# Chinese-Text-Classification

新闻标题文本分类

- ### 1. 背景

基于transformers bert 做了一个文本分类的训练 + 推理流程

### 2. 文件

```shell
  > datasets		数据集文件夹
  > pretrained_model	各种预训练模型文件夹
  > saved_model		微调之后保存的模型文件夹
    train.py		训练文件
    inference.py		推理文件
  ```

### 3. 使用

需要将公开数据集和预训练模型放到指定目录下， 并检查在代码中的位置是否对应

```python
  # 模型保存位置
  SAVE_PATH = 'saved_model/pytorch_model.bin'
  # 预训练模型目录
  BERT = 'pretrained_model/chinese_bert_pytorch'
  model_path = BERT
  # 数据位置
  CLASS_PATH = 'datasets/class.txt'
  TRAIN_PATH = 'datasets/train.txt'
  DEV_PATH = 'datasets/dev.txt'
  TEST_PATH = 'datasets/test.txt'
  ```

训练

```shell
  python train.py
  ```

推理

```shell
  python inference.py
  ```

### 4. 下载

数据集：

- 存放于项目datasets文件夹

预训练模型：

- BERT：https://s3.amazonaws.com/models.huggingface.co/bert/bert-base-chinese.tar.gz
  - BERT-wwm：https://drive.google.com/file/d/1AQitrjbvCWc51SYiLN-cJq4e0WiNN4KY/view
  - BERT-wwm-ext：https://drive.google.com/file/d/1iNeYFhCBJWeUsIlnW_2K6SMwXkM4gLb_/view
  - RoBERTa-wwm-ext：https://drive.google.com/file/d/1eHM3l4fMo6DsQYGmey7UZGiTmQquHw25/view

### 5. 测评

测评指标为acc

参数：batch_size=128，lr=5e-5，maxlen=32，epochs=1（增加epochs可能效果会更好）

| 模型            | dev    | test   |
  | :-------------- | ------ | ------ |
  | BERT            | 0.9351 | 0.9427 |
  | BERT-wwm        | 0.9367 | 0.9395 |
  | BERT-wwm-ext    | 0.9368 | 0.9419 |
  | RoBERTa-wwm-ext | 0.9385 | 0.9450 |

### 6. 参考

- https://github.com/649453932/Bert-Chinese-Text-Classification-Pytorch

Chinese-Text-Classification

Explore Similar Repositories

Was this content helpful?

Recent Feedback

Download README