Joelzhang commited on
Commit
a2043d4
1 Parent(s): cb1df47

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +54 -16
README.md CHANGED
@@ -14,35 +14,73 @@ widget:
14
  - text: "今天心情不好[SEP]今天很开心"
15
 
16
  ---
17
- # Erlangshen-MegatronBert-1.3B-NLI, model (Chinese),one model of [Fengshenbang-LM](https://github.com/IDEA-CCNL/Fengshenbang-LM).
18
- We collect 4 NLI(Natural Language Inference) datasets in the Chinese domain for finetune, with a total of 1014787 samples. Our model is mainly based on [roberta](https://huggingface.co/hfl/chinese-roberta-wwm-ext)
19
 
20
- ## Usage
21
- ```python
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
22
  from transformers import AutoModelForSequenceClassification
23
  from transformers import BertTokenizer
24
  import torch
25
-
26
  tokenizer=BertTokenizer.from_pretrained('IDEA-CCNL/Erlangshen-MegatronBert-1.3B-NLI')
27
  model=AutoModelForSequenceClassification.from_pretrained('IDEA-CCNL/Erlangshen-MegatronBert-1.3B-NLI')
28
-
29
  texta='今天的饭不好吃'
30
  textb='今天心情不好'
31
-
32
  output=model(torch.tensor([tokenizer.encode(texta,textb)]))
33
  print(torch.nn.functional.softmax(output.logits,dim=-1))
34
-
35
  ```
36
- ## Scores on downstream chinese tasks (without any data augmentation)
37
- | Model | cmnli | ocnli | snli |
38
- | :--------: | :-----: | :----: | :-----: |
39
- | Erlangshen-Roberta-110M-NLI | 80.83 | 78.56 | 88.01 |
40
- | Erlangshen-Roberta-330M-NLI | 82.25 | 79.82 | 88 |
41
- | Erlangshen-MegatronBert-1.3B-NLI | 84.52 | 84.17 | 88.67 |
42
 
43
- ## Citation
44
- If you find the resource is useful, please cite the following website in your paper.
 
 
 
 
 
 
 
 
 
 
 
 
45
  ```
 
 
 
 
 
 
46
  @misc{Fengshenbang-LM,
47
  title={Fengshenbang-LM},
48
  author={IDEA-CCNL},
 
14
  - text: "今天心情不好[SEP]今天很开心"
15
 
16
  ---
17
+ # Erlangshen-MegatronBert-1.3B-NLI
 
18
 
19
+ - Github: [Fengshenbang-LM](https://github.com/IDEA-CCNL/Fengshenbang-LM)
20
+ - Docs: [Fengshenbang-Docs](https://fengshenbang-doc.readthedocs.io/)
21
+
22
+ ## 简介 Brief Introduction
23
+
24
+ 2021年登顶FewCLUE和ZeroCLUE的中文BERT,在数个推理任务微调后的版本
25
+
26
+ This is the fine-tuned version of the Chinese BERT model, which topped FewCLUE and ZeroCLUE benchmark in 2021
27
+
28
+ ## 模型分类 Model Taxonomy
29
+
30
+ | 需求 Demand | 任务 Task | 系列 Series | 模型 Model | 参数 Parameter | 额外 Extra |
31
+ | :----: | :----: | :----: | :----: | :----: | :----: |
32
+ | 通用 General | 自然语言理解 NLU | 二郎神 Erlangshen | MegatronBert | 1.3B | NLI |
33
+
34
+ ## 模型信息 Model Information
35
+
36
+ 基于[Erlangshen-MegatronBert-1.3B](https://huggingface.co/IDEA-CCNL/Erlangshen-MegatronBert-1.3B),我们在收集的4个用于finetune的中文领域的NLI(自然语言推理)数据集,总计1014787个样本上微调了一个NLI版本。
37
+
38
+ Based on [Erlangshen-MegatronBert-1.3B](https://huggingface.co/IDEA-CCNL/Erlangshen-MegatronBert-1.3B), we fine-tuned an NLI version on 4 Natural Language Inference (NLI) datasets that are commonly used for fine-tuning in Chinese field, with totaling 1,014,787 samples.
39
+
40
+ ### 下游效果 Performance
41
+
42
+ | 模型 Model | cmnli | ocnli | snli |
43
+ | :--------: | :-----: | :----: | :-----: |
44
+ | Erlangshen-Roberta-110M-NLI | 80.83 | 78.56 | 88.01 |
45
+ | Erlangshen-Roberta-330M-NLI | 82.25 | 79.82 | 88 |
46
+ | Erlangshen-MegatronBert-1.3B-NLI | 84.52 | 84.17 | 88.67 |
47
+
48
+
49
+ ## 使用 Usage
50
+
51
+ ``` python
52
  from transformers import AutoModelForSequenceClassification
53
  from transformers import BertTokenizer
54
  import torch
 
55
  tokenizer=BertTokenizer.from_pretrained('IDEA-CCNL/Erlangshen-MegatronBert-1.3B-NLI')
56
  model=AutoModelForSequenceClassification.from_pretrained('IDEA-CCNL/Erlangshen-MegatronBert-1.3B-NLI')
 
57
  texta='今天的饭不好吃'
58
  textb='今天心情不好'
 
59
  output=model(torch.tensor([tokenizer.encode(texta,textb)]))
60
  print(torch.nn.functional.softmax(output.logits,dim=-1))
 
61
  ```
 
 
 
 
 
 
62
 
63
+ ## 引用 Citation
64
+
65
+ 如果您在您的工作中使用了我们的模型,可以引用我们的[论文](https://arxiv.org/abs/2209.02970):
66
+
67
+ If you are using the resource for your work, please cite the our [paper](https://arxiv.org/abs/2209.02970):
68
+
69
+ ```text
70
+ @article{fengshenbang,
71
+ author = {Junjie Wang and Yuxiang Zhang and Lin Zhang and Ping Yang and Xinyu Gao and Ziwei Wu and Xiaoqun Dong and Junqing He and Jianheng Zhuo and Qi Yang and Yongfeng Huang and Xiayu Li and Yanghan Wu and Junyu Lu and Xinyu Zhu and Weifeng Chen and Ting Han and Kunhao Pan and Rui Wang and Hao Wang and Xiaojun Wu and Zhongshen Zeng and Chongpei Chen and Ruyi Gan and Jiaxing Zhang},
72
+ title = {Fengshenbang 1.0: Being the Foundation of Chinese Cognitive Intelligence},
73
+ journal = {CoRR},
74
+ volume = {abs/2209.02970},
75
+ year = {2022}
76
+ }
77
  ```
78
+
79
+ 也可以引用我们的[网站](https://github.com/IDEA-CCNL/Fengshenbang-LM/):
80
+
81
+ You can also cite our [website](https://github.com/IDEA-CCNL/Fengshenbang-LM/):
82
+
83
+ ```text
84
  @misc{Fengshenbang-LM,
85
  title={Fengshenbang-LM},
86
  author={IDEA-CCNL},