Skip to content

中文长文本分类、短句子分类、多标签分类、两句子相似度(Chinese Text Classification of Keras NLP, multi-label classify, or sentence classify, long or short),字词句向量嵌入层(embeddings)和网络层(graph)构建基类,FastText,TextCNN,CharCNN,TextRNN, RCNN, DCNN, DPCNN, VDCNN, CRNN, Bert, Xlnet, Albert, Attention, DeepMoji, HAN, 胶囊网络-CapsuleNet, Transformer-encode, Seq2seq, SWEM, LEAM, TextGCN

License

Notifications You must be signed in to change notification settings

yongzhuo/Keras-TextClassification

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

PyPI Build Status PyPI_downloads Stars Forks Join the chat at https://gitter.im/yongzhuo/Keras-TextClassification

Install(安装)

pip install Keras-TextClassification
step2: download and unzip the dir of 'data.rar', 地址: 链接https://pan.baidu.com/s/1pIDzGaGXCZ7cjng1XU_kPA   提取码w6ps   压缩包密码: 2022
       cover the dir of data to anaconda, like '/anaconda/3.5.1/envs/tensorflow13/Lib/site-packages/keras_textclassification/data'
step3: goto # Train&Usage(调用) and Predict&Usage(调用)

keras_textclassification(代码主体,未完待续...)

- Electra-fineture(todo)
- Albert-fineture
- Xlnet-fineture
- Bert-fineture
- FastText
- TextCNN
- charCNN
- TextRNN
- TextRCNN
- TextDCNN
- TextDPCNN
- TextVDCNN
- TextCRNN
- DeepMoji
- SelfAttention
- HAN
- CapsuleNet
- Transformer-encode
- SWEM
- LEAM
- TextGCN(todo)

run(运行, 以FastText为例)

- 1. 进入keras_textclassification/m01_FastText目录,
- 2. 训练: 运行 train.py,   例如: python train.py
- 3. 预测: 运行 predict.py, 例如: python predict.py
- 说明: 默认不带pre train的random embedding,训练和验证语料只有100条,完整语料移步下面data查看下载

run(多标签分类/Embedding/test/sample实例)

- bert,word2vec,random样例在test/目录下, 注意word2vec(char or word), random-word,  bert(chinese_L-12_H-768_A-12)未全部加载,需要下载
- multi_multi_class/目录下以text-cnn为例进行多标签分类实例,转化为multi-onehot标签类别,分类则取一定阀值的类
- sentence_similarity/目录下以bert为例进行两个句子文本相似度计算,数据格式如data/sim_webank/目录下所示
- predict_bert_text_cnn.py
- tet_char_bert_embedding.py
- tet_char_bert_embedding.py
- tet_char_xlnet_embedding.py
- tet_char_random_embedding.py
- tet_char_word2vec_embedding.py
- tet_word_random_embedding.py
- tet_word_word2vec_embedding.py

keras_textclassification/data

- 数据下载
  ** github项目中只是上传部分数据,需要的前往链接: https://pan.baidu.com/s/1I3vydhmFEQ9nuPG2fDou8Q 提取码: rket
- baidu_qa_2019(百度qa问答语料,只取title作为分类样本,17个类,有一个是空'',已经压缩上传)
   - baike_qa_train.csv
   - baike_qa_valid.csv
- byte_multi_news(今日头条2018新闻标题多标签语料,1070个标签,fate233爬取, 地址为: [byte_multi_news](https://github.com/fate233/toutiao-multilevel-text-classfication-dataset))
   -labels.csv
   -train.csv
   -valid.csv
- embeddings
   - chinese_L-12_H-768_A-12/(取谷歌预训练好点的模型,已经压缩上传,
                              keras-bert还可以加载百度版ernie(需转换,[https://github.com/ArthurRizar/tensorflow_ernie](https://github.com/ArthurRizar/tensorflow_ernie)),
                              哈工大版bert-wwm(tf框架,[https://github.com/ymcui/Chinese-BERT-wwm](https://github.com/ymcui/Chinese-BERT-wwm))
   - albert_base_zh/(brightmart训练的albert, 地址为https://github.com/brightmart/albert_zh)
   - chinese_xlnet_base_L-12_H-768_A-12/(哈工大预训练的中文xlnet模型[https://github.com/ymcui/Chinese-PreTrained-XLNet],12层)
   - term_char.txt(已经上传, 项目中已全, wiki字典, 还可以用新华字典什么的)
   - term_word.txt(未上传, 项目中只有部分, 可参考词向量的)
   - w2v_model_merge_short.vec(未上传, 项目中只有部分, 词向量, 可以用自己的)
   - w2v_model_wiki_char.vec(已上传百度网盘, 项目中只有部分, 自己训练的维基百科字向量, 可以用自己的)
- model
   - fast_text/预训练模型存放地址

项目说明

    1. 构建了base基类(网络(graph)、向量嵌入(词、字、句子embedding)),后边的具体模型继承它们,代码简单
    1. keras_layers存放一些常用的layer, conf存放项目数据、模型的地址, data存放数据和语料, data_preprocess为数据预处理模块,

模型与论文paper题与地址

参考/感谢

训练简单调用:

from keras_textclassification import train
train(graph='TextCNN', # 必填, 算法名, 可选"ALBERT","BERT","XLNET","FASTTEXT","TEXTCNN","CHARCNN",
                       # "TEXTRNN","RCNN","DCNN","DPCNN","VDCNN","CRNN","DEEPMOJI",
                       # "SELFATTENTION", "HAN","CAPSULE","TRANSFORMER"
     label=17,         # 必填, 类别数, 训练集和测试集合必须一样
     path_train_data=None, # 必填, 训练数据文件, csv格式, 必须含'label,ques'头文件, 详见keras_textclassification/data
     path_dev_data=None, # 必填, 测试数据文件, csv格式, 必须含'label,ques'头文件, 详见keras_textclassification/data
     rate=1,             # 可填, 训练数据选取比例
     hyper_parameters=None) # 可填, json格式, 超参数, 默认embedding为'char','random'

Reference

For citing this work, you can refer to the present GitHub project. For example, with BibTeX:

@misc{Keras-TextClassification,
    howpublished = {\url{https://github.com/yongzhuo/Keras-TextClassification}},
    title = {Keras-TextClassification},
    author = {Yongzhuo Mo},
    publisher = {GitHub},
    year = {2019}
}

*希望对你有所帮助!

About

中文长文本分类、短句子分类、多标签分类、两句子相似度(Chinese Text Classification of Keras NLP, multi-label classify, or sentence classify, long or short),字词句向量嵌入层(embeddings)和网络层(graph)构建基类,FastText,TextCNN,CharCNN,TextRNN, RCNN, DCNN, DPCNN, VDCNN, CRNN, Bert, Xlnet, Albert, Attention, DeepMoji, HAN, 胶囊网络-CapsuleNet, Transformer-encode, Seq2seq, SWEM, LEAM, TextGCN

Topics

Resources

License

Stars

Watchers

Forks

Packages

No packages published

Languages