Skip to content

Latest commit

 

History

History
302 lines (248 loc) · 13.7 KB

README.md

File metadata and controls

302 lines (248 loc) · 13.7 KB

BertWithPretrained

[中文|English]

This project is an implementation of the BERT model and its related downstream tasks based on the PyTorch framework. It also includes a detailed explanation of the BERT model and the principles of each underlying task.

BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding

Before learning to use this project, you need to know the relevant principles of Transformer by these three examples: Translation, Classification, Coupling Generation.

Implementations

Project Structure

  • bert_base_chinesecontains the bert_base_chinese pre-training model and configuration files

  • bert_base_uncased_englishcontains the bert_base_uncased_english pre-training model and configuration files

  • datacontains all datasets used by each downstream task.

    • SingleSentenceClassification is a 15-classes Chinese classification dataset of Toutiao.
    • PairSentenceClassification is the dataset of MNLI (The Multi-Genre Natural Language Inference Corpus).
    • MultipeChoice is the dataset of SWAG.
    • SQuAD is the dataset of SQuAD-V1.1.
    • WikiTextis the Wikipedia English corpus for pre-training.
    • SongCi is SongCi data for Chinese model pre-training
    • ChineseNER is a dataset used to train Chinese Named Entity Recognition.
  • model is the implementation of each module

    • BasicBert contains basic BERT implementation
      • MyTransformer.py self-attention implementation.
      • BertEmbedding.py Input Embedding implementation.
      • BertConfig.py used to import configuration of config.json.
      • Bert.py implementation of bert.
    • DownstreamTasks contains all downstream tasks implementation
      • BertForSentenceClassification.py sentence(s) classification implementation.
      • BertForMultipleChoice.py multiple choice implementation.
      • BertForQuestionAnswering.py question answer (text span) implementation.
      • BertForNSPAndMLM.py NSP and MLM implementation.
      • BertForTokenClassification.py token classification implementation.
  • Task implementation of training and inference for each downstream task

    • TaskForSingleSentenceClassification.py taks of single sentence classification implementation such as sentence classification.
    • TaskForPairSentence.py task of pair sentence classification implementation such as MNLI.
    • TaskForMultipleChoice.py task of multiple choice implementation such as SWAG.
    • TaskForSQuADQuestionAnswering.py task os question answering (text span) implementation such as SQuAD.
    • TaskForPretraining.py tasks of NSP ans MLM implementation.
    • TaskForChineseNER.py task of Chinese Named Entity Recognition implementation.
  • test test cases of each downstream task.

  • utils

    • data_helpers.py is the data preprocessing and dataset building module of each downstream task;
    • log_helper.py is the log printing module.
    • creat_pretraining_data.py used to construct the dataset of BERT pre-training task.

Python Environment

Python 3.6 and packages version

torch==1.5.0
torchtext==0.6.0
torchvision==0.6.0
transformers==4.5.1
numpy==1.19.5
pandas==1.1.5
scikit-learn==0.24.0
tqdm==4.61.0

Usage

Step 1. Download Dataset

Downloading each dataset and the corresponding BERT pretrained model (if empty) and putting it in the corresponding directory. For details, see the README.md file in each data (data) directory.

Step 2. Runing

Going to the Tasks directory and run the model.

2.1 Chinese text classification task

Model structure and data processing:

python TaskForSingleSentenceClassification.py

Result:

-- INFO: Epoch: 0, Batch[0/4186], Train loss :2.862, Train acc: 0.125
-- INFO: Epoch: 0, Batch[10/4186], Train loss :2.084, Train acc: 0.562
-- INFO: Epoch: 0, Batch[20/4186], Train loss :1.136, Train acc: 0.812        
-- INFO: Epoch: 0, Batch[30/4186], Train loss :1.000, Train acc: 0.734
...
-- INFO: Epoch: 0, Batch[4180/4186], Train loss :0.418, Train acc: 0.875
-- INFO: Epoch: 0, Train loss: 0.481, Epoch time = 1123.244s
...
-- INFO: Epoch: 9, Batch[4180/4186], Train loss :0.102, Train acc: 0.984
-- INFO: Epoch: 9, Train loss: 0.100, Epoch time = 1130.071s
-- INFO: Accurcay on val 0.884
-- INFO: Accurcay on val 0.888

2.2 Text Implication

Model structure and data processing:

python TaskForPairSentenceClassification.py

Result:

-- INFO: Epoch: 0, Batch[0/17181], Train loss :1.082, Train acc: 0.438
-- INFO: Epoch: 0, Batch[10/17181], Train loss :1.104, Train acc: 0.438
-- INFO: Epoch: 0, Batch[20/17181], Train loss :1.129, Train acc: 0.250     
-- INFO: Epoch: 0, Batch[30/17181], Train loss :1.063, Train acc: 0.375
...
-- INFO: Epoch: 0, Batch[17180/17181], Train loss :0.367, Train acc: 0.909
-- INFO: Epoch: 0, Train loss: 0.589, Epoch time = 2610.604s
...
-- INFO: Epoch: 9, Batch[0/17181], Train loss :0.064, Train acc: 1.000
-- INFO: Epoch: 9, Train loss: 0.142, Epoch time = 2542.781s
-- INFO: Accurcay on val 0.827
-- INFO: Accurcay on val 0.830

2.3 Multiple Choice (SWAG) Task

Model structure and data processing:

python TaskForMultipleChoice.py

Result:

[2021-11-11 21:32:50] - INFO: Epoch: 0, Batch[0/4597], Train loss :1.433, Train acc: 0.250
[2021-11-11 21:32:58] - INFO: Epoch: 0, Batch[10/4597], Train loss :1.277, Train acc: 0.438
[2021-11-11 21:33:01] - INFO: Epoch: 0, Batch[20/4597], Train loss :1.249, Train acc: 0.438
        ......
[2021-11-11 21:58:34] - INFO: Epoch: 0, Batch[4590/4597], Train loss :0.489, Train acc: 0.875
[2021-11-11 21:58:36] - INFO: Epoch: 0, Batch loss :0.786, Epoch time = 1546.173s
[2021-11-11 21:28:55] - INFO: Epoch: 0, Batch[0/4597], Train loss :1.433, Train acc: 0.250
[2021-11-11 21:30:52] - INFO: He is throwing darts at a wall. A woman, squats alongside flies side to side with his gun.  ## False
[2021-11-11 21:30:52] - INFO: He is throwing darts at a wall. A woman, throws a dart at a dartboard.   ## False
[2021-11-11 21:30:52] - INFO: He is throwing darts at a wall. A woman, collapses and falls to the floor.   ## False
[2021-11-11 21:30:52] - INFO: He is throwing darts at a wall. A woman, is standing next to him.    ## True
[2021-11-11 21:30:52] - INFO: Accuracy on val 0.794

2.4 Question Answering (SQuAD) Task

Model structure and data processing:

python TaskForSQuADQuestionAnswering.py

Result:

[2022-01-02 14:42:17]缓存文件 ~/BertWithPretrained/data/SQuAD/dev-v1_128_384_64.pt 不存在重新处理并缓存!
[2022-01-02 14:42:17] - DEBUG: <<<<<<<<  进入新的example  >>>>>>>>>
[2022-01-02 14:42:17] - DEBUG: ## 正在预处理数据 utils.data_helpers is_training = False
[2022-01-02 14:42:17] - DEBUG: ## 问题 id: 56be5333acb8001400a5030d
[2022-01-02 14:42:17] - DEBUG: ## 原始问题 text: Which performers joined the headliner during the Super Bowl 50 halftime show?
[2022-01-02 14:42:17] - DEBUG: ## 原始描述 text: CBS broadcast Super Bowl 50 in the U.S., and charged an average of $5 million for a  ....
[2022-01-02 14:42:17]- DEBUG: ## 上下文长度为:87, 剩余长度 rest_len 为 : 367
[2022-01-02 14:42:17] - DEBUG: ## input_tokens: ['[CLS]', 'which', 'performers', 'joined', 'the', 'headline', '##r', 'during', 'the', ...]
[2022-01-02 14:42:17] - DEBUG: ## input_ids:[101, 2029, 9567, 2587, 1996, 17653, 2099, 2076, 1996, 3565, 4605, 2753, 22589, 2265, 1029, 102, 6568, ....]
[2022-01-02 14:42:17] - DEBUG: ## segment ids:[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, ...]
[2022-01-02 14:42:17] - DEBUG: ## orig_map:{16: 0, 17: 1, 18: 2, 19: 3, 20: 4, 21: 5, 22: 6, 23: 7, 24: 7, 25: 7, 26: 7, 27: 7, 28: 8, 29: 9, 30: 10,....}
[2022-01-02 14:42:17] - DEBUG: ======================
....
[2022-01-02 15:13:50] - INFO: Epoch:0, Batch[810/7387] Train loss: 0.998, Train acc: 0.708
[2022-01-02 15:13:55] - INFO: Epoch:0, Batch[820/7387] Train loss: 1.130, Train acc: 0.708
[2022-01-02 15:13:59] - INFO: Epoch:0, Batch[830/7387] Train loss: 1.960, Train acc: 0.375
[2022-01-02 15:14:04] - INFO: Epoch:0, Batch[840/7387] Train loss: 1.933, Train acc: 0.542
......
[2022-01-02 15:15:27] - INFO:  ### Quesiotn: [CLS] when was the first university in switzerland founded..
[2022-01-02 15:15:27] - INFO:    ## Predicted answer: 1460
[2022-01-02 15:15:27] - INFO:    ## True answer: 1460
[2022-01-02 15:15:27] - INFO:    ## True answer idx: (tensor(46, tensor(47))
[2022-01-02 15:15:27] - INFO:  ### Quesiotn: [CLS] how many wards in plymouth elect two councillors?
[2022-01-02 15:15:27] - INFO:    ## Predicted answer: 17 of which elect three .....
[2022-01-02 15:15:27] - INFO:    ## True answer: three
[2022-01-02 15:15:27] - INFO:    ## True answer idx: (tensor(25, tensor(25))

运行结束后,data/SQuAD目录中会生成一个名为best_result.json的预测文件,此时只需要切换到该目录下,并运行以下代码即可得到在dev-v1.1.json的测试结果:

python evaluate-v1.1.py dev-v1.1.json best_result.json

"exact_match" : 80.879848628193, "f1": 88.338575234135

2.5 NSL and MLM tasks

Model structure and data processing:

if __name__ == '__main__':
    config = ModelConfig()
    train(config)
    sentences_1 = ["I no longer love her, true, but perhaps I love her.",
                   "Love is so short and oblivion so long."]

    sentences_2 = ["我住长江头,君住长江尾。",
                   "日日思君不见君,共饮长江水。",
                   "此水几时休,此恨何时已。",
                   "只愿君心似我心,定不负相思意。"]
    inference(config, sentences_2, masked=False, language='zh')

Result:

- INFO: ## 成功载入已有模型进行推理……
- INFO:  ### 原始:我住长江头,君住长江尾。
- INFO:   ## 掩盖:我住长江头,[MASK]住长[MASK]尾。
- INFO:   ## 预测:我住长江头,君住长河尾。  
- INFO: ====================
- INFO:  ### 原始:日日思君不见君,共饮长江水。
- INFO:   ## 掩盖:日日思君不[MASK]君,共[MASK]长江水。
- INFO:   ## 预测:日日思君不见君,共饮长江水。
#   ......

2.6 Named Entity Recognition task

Model structure and data processing:

if __name__ == '__main__':
    config = ModelConfig()
    train(config)
    sentences = ['智光拿出石壁拓文为乔峰详述事情始末,乔峰方知自己原本姓萧,乃契丹后族。',
                 '当乔峰问及带头大哥时,却发现智光大师已圆寂。',
                 '乔峰、阿朱相约找最后知情人康敏问完此事后,就到塞外骑马牧羊,再不回来。']
    inference(config, sentences)

Result of training:

- INFO: Epoch: [1/10], Batch[620/1739], Train Loss: 0.115, Train acc: 0.96386
- INFO: Epoch: [1/10], Batch[240/1739], Train Loss: 0.098, Train acc: 0.96466
- INFO: Epoch: [1/10], Batch[660/1739], Train Loss: 0.087, Train acc: 0.96435
......
- INFO:句子在澳大利亚等西方国家改变反倾销政策中对中国的划分后不少欧盟人士也认识到此种划分已背离中国经济迅速发展的现实- INFO: 澳大利亚LOC
- INFO: 中国LOC
- INFO: 欧盟LOC
- INFO: 中国LOC
......
              precision    recall  f1-score   support

           O       1.00      0.99      1.00     97640
       B-ORG       0.86      0.93      0.89       984
       B-LOC       0.94      0.93      0.94      1934
       B-PER       0.97      0.97      0.97       884
       I-ORG       0.90      0.96      0.93      3945
       I-LOC       0.91      0.95      0.93      2556
       I-PER       0.99      0.98      0.98      1714

    accuracy                           0.99    109657
   macro avg       0.94      0.96      0.95    109657
weighted avg       0.99      0.99      0.99    109657

Result of inference:

- INFO: 句子智光拿出石壁拓文为乔峰详述事情始末乔峰方知自己原本姓萧乃契丹后族- INFO: 	智光PER
- INFO: 	乔峰PER
- INFO: 	乔峰PER
- INFO: 	PER
- INFO: 	PER
......