Joint Laboratory of HIT and iFLYTEK Research (HFL)
Update README.md 2b36af9
1 ---
2 language:
3 - zh
4 license: "apache-2.0"
5 ---
6
7 # This model is trained on 180G data, we recommend using this one than the original version.
8
9 ## Chinese ELECTRA
10 Google and Stanford University released a new pre-trained model called ELECTRA, which has a much compact model size and relatively competitive performance compared to BERT and its variants.
11 For further accelerating the research of the Chinese pre-trained model, the Joint Laboratory of HIT and iFLYTEK Research (HFL) has released the Chinese ELECTRA models based on the official code of ELECTRA.
12 ELECTRA-small could reach similar or even higher scores on several NLP tasks with only 1/10 parameters compared to BERT and its variants.
13
14 This project is based on the official code of ELECTRA: [https://github.com/google-research/electra](https://github.com/google-research/electra)
15
16 You may also interested in,
17 - Chinese BERT series: https://github.com/ymcui/Chinese-BERT-wwm
18 - Chinese ELECTRA: https://github.com/ymcui/Chinese-ELECTRA
19 - Chinese XLNet: https://github.com/ymcui/Chinese-XLNet
20 - Knowledge Distillation Toolkit - TextBrewer: https://github.com/airaria/TextBrewer
21
22 More resources by HFL: https://github.com/ymcui/HFL-Anthology
23
24
25 ## Citation
26 If you find our resource or paper is useful, please consider including the following citation in your paper.
27 - https://arxiv.org/abs/2004.13922
28 ```
29 @inproceedings{cui-etal-2020-revisiting,
30 title = "Revisiting Pre-Trained Models for {C}hinese Natural Language Processing",
31 author = "Cui, Yiming and
32 Che, Wanxiang and
33 Liu, Ting and
34 Qin, Bing and
35 Wang, Shijin and
36 Hu, Guoping",
37 booktitle = "Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing: Findings",
38 month = nov,
39 year = "2020",
40 address = "Online",
41 publisher = "Association for Computational Linguistics",
42 url = "https://www.aclweb.org/anthology/2020.findings-emnlp.58",
43 pages = "657--668",
44 }
45 ```