kimsan0622 nazneen commited on
Commit
2668e71
1 Parent(s): 436ee07

model documentation (#3)

Browse files

- model documentation (3d9d378fcc2e6e7c406867a15b529339a6d402e5)


Co-authored-by: Nazneen Rajani <[email protected]>

Files changed (1) hide show
  1. README.md +172 -31
README.md CHANGED
@@ -1,43 +1,184 @@
 
1
  ---
2
- license: apache-2.0
3
  language: ko
 
4
  tags:
5
- - t5
6
- eos_token: "</s>"
7
  widget:
8
- - text: 아버지가 방에 들어가신다.</s>
9
  ---
10
 
11
- # ke-t5 base
12
-
13
- Pretrained T5 Model on Korean and English. See [Github](https://github.com/AIRC-KETI/ke-t5) and [Paper](https://aclanthology.org/2021.findings-emnlp.33/) [Korean paper](https://koreascience.kr/article/CFKO202130060717834.pdf) for more details.
14
-
15
- ## How to use
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
16
 
17
  ```python
18
- from transformers import AutoModel, AutoTokenizer
19
-
20
- model = AutoModel.from_pretrained("KETI-AIR/ke-t5-base-ko")
21
  tokenizer = AutoTokenizer.from_pretrained("KETI-AIR/ke-t5-base-ko")
 
 
 
22
  ```
 
23
 
24
- ## BibTeX entry and citation info
25
 
26
- ```bibtex
27
- @inproceedings{kim-etal-2021-model-cross,
28
- title = "A Model of Cross-Lingual Knowledge-Grounded Response Generation for Open-Domain Dialogue Systems",
29
- author = "Kim, San and
30
- Jang, Jin Yea and
31
- Jung, Minyoung and
32
- Shin, Saim",
33
- booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2021",
34
- month = nov,
35
- year = "2021",
36
- address = "Punta Cana, Dominican Republic",
37
- publisher = "Association for Computational Linguistics",
38
- url = "https://aclanthology.org/2021.findings-emnlp.33",
39
- doi = "10.18653/v1/2021.findings-emnlp.33",
40
- pages = "352--365",
41
- abstract = "Research on open-domain dialogue systems that allow free topics is challenging in the field of natural language processing (NLP). The performance of the dialogue system has been improved recently by the method utilizing dialogue-related knowledge; however, non-English dialogue systems suffer from reproducing the performance of English dialogue systems because securing knowledge in the same language with the dialogue system is relatively difficult. Through experiments with a Korean dialogue system, this paper proves that the performance of a non-English dialogue system can be improved by utilizing English knowledge, highlighting the system uses cross-lingual knowledge. For the experiments, we 1) constructed a Korean version of the Wizard of Wikipedia dataset, 2) built Korean-English T5 (KE-T5), a language model pre-trained with Korean and English corpus, and 3) developed a knowledge-grounded Korean dialogue model based on KE-T5. We observed the performance improvement in the open-domain Korean dialogue model even only English knowledge was given. The experimental results showed that the knowledge inherent in cross-lingual language models can be helpful for generating responses in open dialogue systems.",
42
- }
43
- ```
 
1
+
2
  ---
 
3
  language: ko
4
+ license: apache-2.0
5
  tags:
6
+ - t5
7
+ eos_token: </s>
8
  widget:
9
+ - text: 아버지가 방에 들어가신다.</s>
10
  ---
11
 
12
+ # Model Card for ke-t5-base-ko
13
+
14
+
15
+ # Model Details
16
+
17
+ ## Model Description
18
+
19
+
20
+ - **Developed by:** Korea Electronics Technology Institute Artificial Intelligence Research Center
21
+ - **Shared by [Optional]:** More information needed
22
+ - **Model type:** Text2Text Generation
23
+ - **Language(s) (NLP):** More information needed
24
+ - **License:** More information needed
25
+ - **Related Models:**
26
+ - **Parent Model:** T5
27
+ - **Resources for more information:**
28
+ - [Associated Paper](https://jmlr.org/papers/volume21/20-074/20-074.pdf)
29
+ - [Blog Post](https://ai.googleblog.com/2020/02/exploring-transfer-learning-with-t5.html)
30
+
31
+
32
+ # Uses
33
+
34
+
35
+ ## Direct Use
36
+
37
+ This model can be used for the task of Text2Text Generation
38
+
39
+ ## Downstream Use [Optional]
40
+
41
+ More information needed
42
+
43
+ ## Out-of-Scope Use
44
+
45
+ The model should not be used to intentionally create hostile or alienating environments for people.
46
+
47
+ # Bias, Risks, and Limitations
48
+
49
+ Significant research has explored bias and fairness issues with language models (see, e.g., [Sheng et al. (2021)](https://aclanthology.org/2021.acl-long.330.pdf) and [Bender et al. (2021)](https://dl.acm.org/doi/pdf/10.1145/3442188.3445922)). Predictions generated by the model may include disturbing and harmful stereotypes across protected classes; identity characteristics; and sensitive, social, and occupational groups.
50
+
51
+
52
+ ## Recommendations
53
+
54
+ Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
55
+
56
+
57
+ # Training Details
58
+
59
+ ## Training Data
60
+
61
+ The model is pre-trained on the [Colossal Clean Crawled Corpus (C4)](https://www.tensorflow.org/datasets/catalog/c4), which was developed and released in the context of the same [research paper](https://jmlr.org/papers/volume21/20-074/20-074.pdf) as T5.
62
+
63
+ The model was pre-trained on a on a **multi-task mixture of unsupervised (1.) and supervised tasks (2.)**.
64
+
65
+ See the [t5-base model card](https://huggingface.co/t5-base?text=My+name+is+Wolfgang+and+I+live+in+Berlin) for further information.
66
+
67
+ ## Training Procedure
68
+
69
+
70
+ ### Preprocessing
71
+
72
+ More information needed
73
+
74
+ ### Speeds, Sizes, Times
75
+
76
+ More information needed
77
+
78
+ # Evaluation
79
+
80
+
81
+ ## Testing Data, Factors & Metrics
82
+
83
+ ### Testing Data
84
+
85
+ More information needed
86
+
87
+ ### Factors
88
+
89
+
90
+ ### Metrics
91
+
92
+ More information needed
93
+ ## Results
94
+
95
+ More information needed
96
+
97
+ # Model Examination
98
+
99
+ More information needed
100
+
101
+ # Environmental Impact
102
+
103
+
104
+ Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
105
+
106
+ - **Hardware Type:** More information needed
107
+ - **Hours used:** More information needed
108
+ - **Cloud Provider:** More information needed
109
+ - **Compute Region:** More information needed
110
+ - **Carbon Emitted:** More information needed
111
+
112
+ # Technical Specifications [optional]
113
+
114
+ ## Model Architecture and Objective
115
+
116
+ More information needed
117
+
118
+ ## Compute Infrastructure
119
+
120
+ More information needed
121
+
122
+ ### Hardware
123
+
124
+ More information needed
125
+
126
+ ### Software
127
+ More information needed
128
+
129
+ # Citation
130
+
131
+ **BibTeX:**
132
+ ```bibtex
133
+ @article{2020t5,
134
+ author = {Colin Raffel and Noam Shazeer and Adam Roberts and Katherine Lee and Sharan Narang and Michael Matena and Yanqi Zhou and Wei Li and Peter J. Liu},
135
+ title = {Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer},
136
+ journal = {Journal of Machine Learning Research},
137
+ year = {2020},
138
+ volume = {21},
139
+ number = {140},
140
+ pages = {1-67},
141
+ url = {http://jmlr.org/papers/v21/20-074.html}
142
+ }
143
+ ```
144
+
145
+
146
+ **APA:**
147
+ ```
148
+ - Raffel, C., Shazeer, N., Roberts, A., Lee, K., Narang, S., Matena, M., ... & Liu, P. J. (2020). Exploring the limits of transfer learning with a unified text-to-text transformer. J. Mach. Learn. Res., 21(140), 1-67.
149
+ ```
150
+
151
+ # Glossary [optional]
152
+ More information needed
153
+
154
+ # More Information [optional]
155
+
156
+ More information needed
157
+
158
+ # Model Card Authors [optional]
159
+
160
+
161
+ Korea Electronics Technology Institute Artificial Intelligence Research Center in collaboration with Ezi Ozoani and the Hugging Face team
162
+
163
+ # Model Card Contact
164
+
165
+ More information needed
166
+
167
+ # How to Get Started with the Model
168
+
169
+ Use the code below to get started with the model.
170
+
171
+ <details>
172
+ <summary> Click to expand </summary>
173
 
174
  ```python
175
+ from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
176
+
 
177
  tokenizer = AutoTokenizer.from_pretrained("KETI-AIR/ke-t5-base-ko")
178
+
179
+ model = AutoModelForSeq2SeqLM.from_pretrained("KETI-AIR/ke-t5-base-ko")
180
+
181
  ```
182
+ </details>
183
 
 
184