https://github.com/bojone/BERT-whitening

简单的向量白化改善句向量质量
https://github.com/bojone/BERT-whitening

Last synced: about 1 month ago
JSON representation

简单的向量白化改善句向量质量

• Host: GitHub
• URL: https://github.com/bojone/BERT-whitening
• Owner: bojone
• Created: 2021-01-11T02:51:18.000Z (almost 4 years ago)
• Default Branch: main
• Last Pushed: 2021-06-17T03:40:26.000Z (over 3 years ago)
• Last Synced: 2024-08-02T16:55:44.082Z (5 months ago)
• Language: Python
• Homepage:
• Size: 85 KB
• Stars: 478
• Watchers: 6
• Forks: 66
• Open Issues: 5
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

• StarryDivineSky - bojone/BERT-whitening - flow的效果。 (BERT优化 / 大语言对话模型及数据)

README

        
# BERT-whitening

通过简单的向量白化来改善句向量质量，可以媲美甚至超过BERT-flow的效果。

## 介绍

- 博客：https://kexue.fm/archives/8321

- 博客：https://kexue.fm/archives/8069

- 论文：[《Whitening Sentence Representations for Better Semantics and Faster Retrieval》](https://arxiv.org/abs/2103.15316)

## 使用

- [eng](https://github.com/bojone/BERT-whitening/tree/main/eng): BERT-whitening在常见英文数据集上的测试；

- [chn](https://github.com/bojone/BERT-whitening/tree/main/chn): BERT-whitening在常见中文数据集上的测试；

- [demo.py](https://github.com/bojone/BERT-whitening/blob/main/demo.py): 早期的简单Demo。

测试环境：tensorflow 1.14 + keras 2.3.1 + bert4keras 0.10.5，如果在其他环境组合下报错，请根据错误信息自行调整代码。

## 下载

文件合集：链接: https://pan.baidu.com/s/1qv7HuYN3bNvEzPrlnlJqEQ 提取码: vbc3

## 引用

```

@article{su2021whitening,

  title={Whitening Sentence Representations for Better Semantics and Faster Retrieval},

  author={Su, Jianlin and Cao, Jiarun and Liu, Weijie and Ou, Yangyiwen},

  journal={arXiv preprint arXiv:2103.15316},

  year={2021}

}

```

## 其他

PyTorch版：https://github.com/autoliuweijie/BERT-whitening-pytorch

## 交流

QQ交流群：808623966，微信群请加机器人微信号spaces_ac_cn