File size: 12,678 Bytes
e4abeaa 5b4ca19 e4abeaa 5b4ca19 e4abeaa 5b4ca19 e4abeaa 5b4ca19 e4abeaa 5b4ca19 e4abeaa 5b4ca19 e4abeaa 5b4ca19 e4abeaa 5b4ca19 e4abeaa 5b4ca19 e4abeaa 5b4ca19 e4abeaa 5b4ca19 e4abeaa 5b4ca19 e4abeaa 5b4ca19 e4abeaa 5b4ca19 e4abeaa 5b4ca19 e4abeaa 5b4ca19 e4abeaa 5b4ca19 e4abeaa 5b4ca19 e4abeaa 5b4ca19 e4abeaa 5b4ca19 e4abeaa 5b4ca19 e4abeaa 5b4ca19 e4abeaa 5b4ca19 e4abeaa 5b4ca19 e4abeaa 5b4ca19 e4abeaa 5b4ca19 e4abeaa 5b4ca19 e4abeaa 5b4ca19 e4abeaa 5b4ca19 e4abeaa 5b4ca19 e4abeaa 5b4ca19 e4abeaa 5b4ca19 e4abeaa 5b4ca19 e4abeaa 5b4ca19 e4abeaa 5b4ca19 e4abeaa 5b4ca19 e4abeaa 5b4ca19 e4abeaa 5b4ca19 e4abeaa 5b4ca19 e4abeaa 5b4ca19 e4abeaa 5b4ca19 e4abeaa 5b4ca19 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 |
---
language:
- ko
- en
license: mit
tags:
- electra
- korean
---
# Model Card for KcELECTRA: Korean comments ELECTRA
# Model Details
## Model Description
** Updates on 2022.10.08 **
- KcELECTRA-base-v2022 (๊ตฌ v2022-dev) ๋ชจ๋ธ ์ด๋ฆ์ด ๋ณ๊ฒฝ๋์์ต๋๋ค.
- ์ ๋ชจ๋ธ์ ์ธ๋ถ ์ค์ฝ์ด๋ฅผ ์ถ๊ฐํ์์ต๋๋ค.
- ๊ธฐ์กด KcELECTRA-base(v2021) ๋๋น ๋๋ถ๋ถ์ downstream task์์ ~1%p ์์ค์ ์ฑ๋ฅ ํฅ์์ด ์์ต๋๋ค.
---
๊ณต๊ฐ๋ ํ๊ตญ์ด Transformer ๊ณ์ด ๋ชจ๋ธ๋ค์ ๋๋ถ๋ถ ํ๊ตญ์ด ์ํค, ๋ด์ค ๊ธฐ์ฌ, ์ฑ
๋ฑ ์ ์ ์ ๋ ๋ฐ์ดํฐ๋ฅผ ๊ธฐ๋ฐ์ผ๋ก ํ์ตํ ๋ชจ๋ธ์
๋๋ค. ํํธ, ์ค์ ๋ก NSMC์ ๊ฐ์ User-Generated Noisy text domain ๋ฐ์ดํฐ์
์ ์ ์ ๋์ง ์์๊ณ ๊ตฌ์ด์ฒด ํน์ง์ ์ ์กฐ์ด๊ฐ ๋ง์ผ๋ฉฐ, ์คํ์ ๋ฑ ๊ณต์์ ์ธ ๊ธ์ฐ๊ธฐ์์ ๋ํ๋์ง ์๋ ํํ๋ค์ด ๋น๋ฒํ๊ฒ ๋ฑ์ฅํฉ๋๋ค.
KcELECTRA๋ ์์ ๊ฐ์ ํน์ฑ์ ๋ฐ์ดํฐ์
์ ์ ์ฉํ๊ธฐ ์ํด, ๋ค์ด๋ฒ ๋ด์ค์์ ๋๊ธ๊ณผ ๋๋๊ธ์ ์์งํด, ํ ํฌ๋์ด์ ์ ELECTRA๋ชจ๋ธ์ ์ฒ์๋ถํฐ ํ์ตํ Pretrained ELECTRA ๋ชจ๋ธ์
๋๋ค.
๊ธฐ์กด KcBERT ๋๋น ๋ฐ์ดํฐ์
์ฆ๊ฐ ๋ฐ vocab ํ์ฅ์ ํตํด ์๋นํ ์์ค์ผ๋ก ์ฑ๋ฅ์ด ํฅ์๋์์ต๋๋ค.
KcELECTRA๋ Huggingface์ Transformers ๋ผ์ด๋ธ๋ฌ๋ฆฌ๋ฅผ ํตํด ๊ฐํธํ ๋ถ๋ฌ์ ์ฌ์ฉํ ์ ์์ต๋๋ค. (๋ณ๋์ ํ์ผ ๋ค์ด๋ก๋๊ฐ ํ์ํ์ง ์์ต๋๋ค.)
- **Developed by:** Junbum Lee
- **Shared by [Optional]:** Hugging Face
- **Model type:** electra
- **Language(s) (NLP):** en
- **License:** MIT
- **Related Models:**
- **Parent Model:** Electra
- **Resources for more information:**
- [GitHub Repo](https://github.com/Beomi/KcBERT-finetune )
- [Model Space](https://huggingface.co/spaces/BeMerciless/korean_malicious_comment)
- [Blog Post](ttps://monologg.kr/categories/NLP/ELECTRA/)
# Uses
## Direct Use
This model can be used for the task of
## Downstream Use [Optional]
More information needed
## Out-of-Scope Use
The model should not be used to intentionally create hostile or alienating environments for people.
# Bias, Risks, and Limitations
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.
## Recommendations
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
# Training Details
## Training Data
ํ์ต ๋ฐ์ดํฐ๋ 2019.01.01 ~ 2021.03.09 ์ฌ์ด์ ์์ฑ๋ **๋๊ธ ๋ง์ ๋ด์ค/ํน์ ์ ์ฒด ๋ด์ค** ๊ธฐ์ฌ๋ค์ **๋๊ธ๊ณผ ๋๋๊ธ**์ ๋ชจ๋ ์์งํ ๋ฐ์ดํฐ์
๋๋ค.
๋ฐ์ดํฐ ์ฌ์ด์ฆ๋ ํ
์คํธ๋ง ์ถ์ถ์ **์ฝ 17.3GB์ด๋ฉฐ, 1์ต8์ฒ๋ง๊ฐ ์ด์์ ๋ฌธ์ฅ**์ผ๋ก ์ด๋ค์ ธ ์์ต๋๋ค.
> KcBERT๋ 2019.01-2020.06์ ํ
์คํธ๋ก, ์ ์ ํ ์ฝ 9์ฒ๋ง๊ฐ ๋ฌธ์ฅ์ผ๋ก ํ์ต์ ์งํํ์ต๋๋ค.
#### Finetune Samples
- NSMC with PyTorch-Lightning 1.3.0, GPU, Colab <a href="https://colab.research.google.com/drive/1Hh63kIBAiBw3Hho--BvfdUWLu-ysMFF0?usp=sharing">
<img src="https://colab.research.google.com/assets/colab-badge.svg" alt="Open In Colab"/>
</a>
## Training Procedure
### Preprocessing
PLM ํ์ต์ ์ํด์ ์ ์ฒ๋ฆฌ๋ฅผ ์งํํ ๊ณผ์ ์ ๋ค์๊ณผ ๊ฐ์ต๋๋ค.
1. ํ๊ธ ๋ฐ ์์ด, ํน์๋ฌธ์, ๊ทธ๋ฆฌ๊ณ ์ด๋ชจ์ง(๐ฅณ)๊น์ง!
์ ๊ทํํ์์ ํตํด ํ๊ธ, ์์ด, ํน์๋ฌธ์๋ฅผ ํฌํจํด Emoji๊น์ง ํ์ต ๋์์ ํฌํจํ์ต๋๋ค.
ํํธ, ํ๊ธ ๋ฒ์๋ฅผ `ใฑ-ใ
๊ฐ-ํฃ` ์ผ๋ก ์ง์ ํด `ใฑ-ํฃ` ๋ด์ ํ์๋ฅผ ์ ์ธํ์ต๋๋ค.
2. ๋๊ธ ๋ด ์ค๋ณต ๋ฌธ์์ด ์ถ์ฝ
`ใ
ใ
ใ
ใ
ใ
`์ ๊ฐ์ด ์ค๋ณต๋ ๊ธ์๋ฅผ `ใ
ใ
`์ ๊ฐ์ ๊ฒ์ผ๋ก ํฉ์ณค์ต๋๋ค.
3. Cased Model
KcBERT๋ ์๋ฌธ์ ๋ํด์๋ ๋์๋ฌธ์๋ฅผ ์ ์งํ๋ Cased model์
๋๋ค.
4. ๊ธ์ ๋จ์ 10๊ธ์ ์ดํ ์ ๊ฑฐ
10๊ธ์ ๋ฏธ๋ง์ ํ
์คํธ๋ ๋จ์ผ ๋จ์ด๋ก ์ด๋ค์ง ๊ฒฝ์ฐ๊ฐ ๋ง์ ํด๋น ๋ถ๋ถ์ ์ ์ธํ์ต๋๋ค.
5. ์ค๋ณต ์ ๊ฑฐ
์ค๋ณต์ ์ผ๋ก ์ฐ์ธ ๋๊ธ์ ์ ๊ฑฐํ๊ธฐ ์ํด ์์ ํ ์ผ์นํ๋ ์ค๋ณต ๋๊ธ์ ํ๋๋ก ํฉ์ณค์ต๋๋ค.
6. `OOO` ์ ๊ฑฐ
๋ค์ด๋ฒ ๋๊ธ์ ๊ฒฝ์ฐ, ๋น์์ด๋ ์์ฒด ํํฐ๋ง์ ํตํด `OOO` ๋ก ํ์ํฉ๋๋ค. ์ด ๋ถ๋ถ์ ๊ณต๋ฐฑ์ผ๋ก ์ ๊ฑฐํ์์ต๋๋ค.
### Speeds, Sizes, Times
More information needed
# Evaluation
## Testing Data, Factors & Metrics
### Testing Data
#### Cleaned Data
- KcBERT ์ธ ์ถ๊ฐ ๋ฐ์ดํฐ๋ ์ ๋ฆฌ ํ ๊ณต๊ฐ ์์ ์
๋๋ค.
### Factors
### Metrics
More information needed
## Results
(100k step๋ณ Checkpoint๋ฅผ ํตํด ์ฑ๋ฅ ํ๊ฐ๋ฅผ ์งํํ์์ต๋๋ค. ํด๋น ๋ถ๋ถ์ `KcBERT-finetune` repo๋ฅผ ์ฐธ๊ณ ํด์ฃผ์ธ์.)
๋ชจ๋ธ ํ์ต Loss๋ Step์ ๋ฐ๋ผ ์ด๊ธฐ 100-200k ์ฌ์ด์ ๊ธ๊ฒฉํ Loss๊ฐ ์ค์ด๋ค๋ค ํ์ต ์ข
๋ฃ๊น์ง๋ ์ง์์ ์ผ๋ก loss๊ฐ ๊ฐ์ํ๋ ๊ฒ์ ๋ณผ ์ ์์ต๋๋ค.
![KcELECTRA-base Pretrain Loss](https://cdn.jsdelivr.net/gh/beomi/blog-img@master/2021/04/07/image-20210407201231133.png)
### KcELECTRA Pretrain Step๋ณ Downstream task ์ฑ๋ฅ ๋น๊ต
> ๐ก ์๋ ํ๋ ์ ์ฒด ckpt๊ฐ ์๋ ์ผ๋ถ์ ๋ํด์๋ง ํ
์คํธ๋ฅผ ์งํํ ๊ฒฐ๊ณผ์
๋๋ค.
![KcELECTRA Pretrain Step๋ณ Downstream task ์ฑ๋ฅ ๋น๊ต](https://cdn.jsdelivr.net/gh/beomi/blog-img@master/2021/04/07/image-20210407215557039.png)
- ์์ ๊ฐ์ด KcBERT-base, KcBERT-large ๋๋น **๋ชจ๋ ๋ฐ์ดํฐ์
์ ๋ํด** KcELECTRA-base๊ฐ ๋ ๋์ ์ฑ๋ฅ์ ๋ณด์
๋๋ค.
- KcELECTRA pretrain์์๋ Train step์ด ๋์ด๊ฐ์ ๋ฐ๋ผ ์ ์ง์ ์ผ๋ก ์ฑ๋ฅ์ด ํฅ์๋๋ ๊ฒ์ ๋ณผ ์ ์์ต๋๋ค.
\***config์ ์ธํ
์ ๊ทธ๋๋ก ํ์ฌ ๋๋ฆฐ ๊ฒฐ๊ณผ์ด๋ฉฐ, hyperparameter tuning์ ์ถ๊ฐ์ ์ผ๋ก ํ ์ ๋ ์ข์ ์ฑ๋ฅ์ด ๋์ฌ ์ ์์ต๋๋ค.**
| | Size<br/>(์ฉ๋) | **NSMC**<br/>(acc) | **Naver NER**<br/>(F1) | **PAWS**<br/>(acc) | **KorNLI**<br/>(acc) | **KorSTS**<br/>(spearman) | **Question Pair**<br/>(acc) | **KorQuaD (Dev)**<br/>(EM/F1) |
| :----------------- | :-------------: | :----------------: | :--------------------: | :----------------: | :------------------: | :-----------------------: | :-------------------------: | :---------------------------: |
| **KcELECTRA-base-v2022** | 475M | **91.97** | 87.35 | 76.50 | 82.12 | 83.67 | 95.12 | 69.00 / 90.40 |
| **KcELECTRA-base** | 475M | 91.71 | 86.90 | 74.80 | 81.65 | 82.65 | **95.78** | 70.60 / 90.11 |
| KcBERT-Base | 417M | 89.62 | 84.34 | 66.95 | 74.85 | 75.57 | 93.93 | 60.25 / 84.39 |
| KcBERT-Large | 1.2G | 90.68 | 85.53 | 70.15 | 76.99 | 77.49 | 94.06 | 62.16 / 86.64 |
| KoBERT | 351M | 89.63 | 86.11 | 80.65 | 79.00 | 79.64 | 93.93 | 52.81 / 80.27 |
| XLM-Roberta-Base | 1.03G | 89.49 | 86.26 | 82.95 | 79.92 | 79.09 | 93.53 | 64.70 / 88.94 |
| HanBERT | 614M | 90.16 | 87.31 | 82.40 | 80.89 | 83.33 | 94.19 | 78.74 / 92.02 |
| KoELECTRA-Base | 423M | 90.21 | 86.87 | 81.90 | 80.85 | 83.21 | 94.20 | 61.10 / 89.59 |
| KoELECTRA-Base-v2 | 423M | 89.70 | 87.02 | 83.90 | 80.61 | 84.30 | 94.72 | 84.34 / 92.58 |
| KoELECTRA-Base-v3 | 423M | 90.63 | **88.11** | **84.45** | **82.24** | **85.53** | 95.25 | **84.83 / 93.45** |
| DistilKoBERT | 108M | 88.41 | 84.13 | 62.55 | 70.55 | 73.21 | 92.48 | 54.12 / 77.80 |
# Model Examination
More information needed
# Environmental Impact
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).
- **Hardware Type:** TPU `v3-8`
- **Hours used:** 240 (10 days)
- **Cloud Provider:** More information needed
- **Compute Region:** More information needed
- **Carbon Emitted:** More information needed
# Technical Specifications [optional]
## Model Architecture and Objective
More information needed
## Compute Infrastructure
More information needed
### Hardware
TPU `v3-8` ์ ์ด์ฉํด ์ฝ 10์ผ ํ์ต์ ์งํํ๊ณ , ํ์ฌ Huggingface์ ๊ณต๊ฐ๋ ๋ชจ๋ธ์ 848k step์ ํ์ตํ ๋ชจ๋ธ weight๊ฐ ์
๋ก๋ ๋์ด์์ต๋๋ค.
### Software
- `pytorch ~= 1.8.0`
- `transformers ~= 4.11.3`
- `emoji ~= 0.6.0`
- `soynlp ~= 0.0.493`
# Citation
**BibTeX:**
```
@misc{lee2021kcelectra,
author = {Junbum Lee},
title = {KcELECTRA: Korean comments ELECTRA},
year = {2021},
publisher = {GitHub},
journal = {GitHub repository},
howpublished = {\url{https://github.com/Beomi/KcELECTRA}}
}
```
๋
ผ๋ฌธ์ ํตํ ์ฌ์ฉ ์ธ์๋ MIT ๋ผ์ด์ผ์ค๋ฅผ ํ๊ธฐํด์ฃผ์ธ์. โบ๏ธ
# Glossary [optional]
More information needed
# More Information [optional]
```
๐ก NOTE ๐ก
General Corpus๋ก ํ์ตํ KoELECTRA๊ฐ ๋ณดํธ์ ์ธ task์์๋ ์ฑ๋ฅ์ด ๋ ์ ๋์ฌ ๊ฐ๋ฅ์ฑ์ด ๋์ต๋๋ค.
KcBERT/KcELECTRA๋ User genrated, Noisy text์ ๋ํด์ ๋ณด๋ค ์ ๋์ํ๋ PLM์
๋๋ค.
```
## Acknowledgement
KcELECTRA Model์ ํ์ตํ๋ GCP/TPU ํ๊ฒฝ์ [TFRC](https://www.tensorflow.org/tfrc?hl=ko) ํ๋ก๊ทธ๋จ์ ์ง์์ ๋ฐ์์ต๋๋ค.
๋ชจ๋ธ ํ์ต ๊ณผ์ ์์ ๋ง์ ์กฐ์ธ์ ์ฃผ์ [Monologg](https://github.com/monologg/) ๋ ๊ฐ์ฌํฉ๋๋ค :)
### Github Repos
- [KcBERT by Beomi](https://github.com/Beomi/KcBERT)
- [BERT by Google](https://github.com/google-research/bert)
- [KoBERT by SKT](https://github.com/SKTBrain/KoBERT)
- [KoELECTRA by Monologg](https://github.com/monologg/KoELECTRA/)
- [Transformers by Huggingface](https://github.com/huggingface/transformers)
- [Tokenizers by Hugginface](https://github.com/huggingface/tokenizers)
- [ELECTRA train code by KLUE](https://github.com/KLUE-benchmark/KLUE-ELECTRA)
# Model Card Authors [optional]
Junbum Lee in collaboration with Ezi Ozoani and the Hugging Face team
# Model Card Contact
More information needed
# How to Get Started with the Model
Use the code below to get started with the model.
<details>
<summary> Click to expand </summary>
```bash
pip install soynlp emoji
```
์๋ `clean` ํจ์๋ฅผ Text data์ ์ฌ์ฉํด์ฃผ์ธ์.
```python
import re
import emoji
from soynlp.normalizer import repeat_normalize
emojis = ''.join(emoji.UNICODE_EMOJI.keys())
pattern = re.compile(f'[^ .,?!/@$%~๏ผ
ยทโผ()\x00-\x7Fใฑ-ใ
ฃ๊ฐ-ํฃ{emojis}]+')
url_pattern = re.compile(
r'https?:\/\/(www\.)?[-a-zA-Z0-9@:%._\+~#=]{1,256}\.[a-zA-Z0-9()]{1,6}\b([-a-zA-Z0-9()@:%_\+.~#?&//=]*)')
import re
import emoji
from soynlp.normalizer import repeat_normalize
pattern = re.compile(f'[^ .,?!/@$%~๏ผ
ยทโผ()\x00-\x7Fใฑ-ใ
ฃ๊ฐ-ํฃ]+')
url_pattern = re.compile(
r'https?:\/\/(www\.)?[-a-zA-Z0-9@:%._\+~#=]{1,256}\.[a-zA-Z0-9()]{1,6}\b([-a-zA-Z0-9()@:%_\+.~#?&//=]*)')
def clean(x):
x = pattern.sub(' ', x)
x = emoji.replace_emoji(x, replace='') #emoji ์ญ์
x = url_pattern.sub('', x)
x = x.strip()
x = repeat_normalize(x, num_repeats=2)
return x
```
</details>
|