ko-barTNumText / README.md
lIlBrother's picture
Update: ์ตœ์ข… ์™„๋ฃŒ ๋ชจ๋ธ์— ๋Œ€ํ•œ README ํ™•์ •
f5442ab
---
language:
- ko # Example: fr
license: apache-2.0 # Example: apache-2.0 or any license from https://hf.co/docs/hub/repositories-licenses
library_name: transformers # Optional. Example: keras or any library from https://github.com/huggingface/hub-docs/blob/main/js/src/lib/interfaces/Libraries.ts
tags:
- text2text-generation # Example: audio
datasets:
- aihub # Example: common_voice. Use dataset id from https://hf.co/datasets
metrics:
- bleu # Example: wer. Use metric id from https://hf.co/metrics
- rouge
# Optional. Add this if you want to encode your eval results in a structured way.
model-index:
- name: ko-barTNumText
results:
- task:
type: text2text-generation # Required. Example: automatic-speech-recognition
name: text2text-generation # Optional. Example: Speech Recognition
metrics:
- type: bleu # Required. Example: wer. Use metric id from https://hf.co/metrics
value: 0.9313276940897475 # Required. Example: 20.90
name: eval_bleu # Optional. Example: Test WER
verified: false # Optional. If true, indicates that evaluation was generated by Hugging Face (vs. self-reported).
- type: rouge1 # Required. Example: wer. Use metric id from https://hf.co/metrics
value: 0.9607081256861959 # Required. Example: 20.90
name: eval_rouge1 # Optional. Example: Test WER
verified: false # Optional. If true, indicates that evaluation was generated by Hugging Face (vs. self-reported).
- type: rouge2 # Required. Example: wer. Use metric id from https://hf.co/metrics
value: 0.9394649136169404 # Required. Example: 20.90
name: eval_rouge2 # Optional. Example: Test WER
verified: false # Optional. If true, indicates that evaluation was generated by Hugging Face (vs. self-reported).
- type: rougeL # Required. Example: wer. Use metric id from https://hf.co/metrics
value: 0.9605735834651536 # Required. Example: 20.90
name: eval_rougeL # Optional. Example: Test WER
verified: false # Optional. If true, indicates that evaluation was generated by Hugging Face (vs. self-reported).
- type: rougeLsum # Required. Example: wer. Use metric id from https://hf.co/metrics
value: 0.9605993760190767 # Required. Example: 20.90
name: eval_rougeLsum # Optional. Example: Test WER
verified: false # Optional. If true, indicates that evaluation was generated by Hugging Face (vs. self-reported).
---
# ko-barTNumText(TNT Model๐Ÿงจ): Try Number To Korean Reading(์ˆซ์ž๋ฅผ ํ•œ๊ธ€๋กœ ๋ฐ”๊พธ๋Š” ๋ชจ๋ธ)
## Table of Contents
- [ko-barTNumText(TNT Model๐Ÿงจ): Try Number To Korean Reading(์ˆซ์ž๋ฅผ ํ•œ๊ธ€๋กœ ๋ฐ”๊พธ๋Š” ๋ชจ๋ธ)](#ko-bartnumtexttnt-model-try-number-to-korean-reading์ˆซ์ž๋ฅผ-ํ•œ๊ธ€๋กœ-๋ฐ”๊พธ๋Š”-๋ชจ๋ธ)
- [Table of Contents](#table-of-contents)
- [Model Details](#model-details)
- [Uses](#uses)
- [Evaluation](#evaluation)
- [How to Get Started With the Model](#how-to-get-started-with-the-model)
## Model Details
- **Model Description:**
๋ญ”๊ฐ€ ์ฐพ์•„๋ด๋„ ๋ชจ๋ธ์ด๋‚˜ ์•Œ๊ณ ๋ฆฌ์ฆ˜์ด ๋”ฑํžˆ ์—†์–ด์„œ ๋งŒ๋“ค์–ด๋ณธ ๋ชจ๋ธ์ž…๋‹ˆ๋‹ค. <br />
BartForConditionalGeneration Fine-Tuning Model For Number To Korean <br />
BartForConditionalGeneration์œผ๋กœ ํŒŒ์ธํŠœ๋‹ํ•œ, ์ˆซ์ž๋ฅผ ํ•œ๊ธ€๋กœ ๋ณ€ํ™˜ํ•˜๋Š” Task ์ž…๋‹ˆ๋‹ค. <br />
- Dataset use [Korea aihub](https://aihub.or.kr/aihubdata/data/list.do?currMenu=115&topMenu=100&srchDataRealmCode=REALM002&srchDataTy=DATA004) <br />
I can't open my fine-tuning datasets for my private issue <br />
๋ฐ์ดํ„ฐ์…‹์€ Korea aihub์—์„œ ๋ฐ›์•„์„œ ์‚ฌ์šฉํ•˜์˜€์œผ๋ฉฐ, ํŒŒ์ธํŠœ๋‹์— ์‚ฌ์šฉ๋œ ๋ชจ๋“  ๋ฐ์ดํ„ฐ๋ฅผ ์‚ฌ์ •์ƒ ๊ณต๊ฐœํ•ด๋“œ๋ฆด ์ˆ˜๋Š” ์—†์Šต๋‹ˆ๋‹ค. <br />
- Korea aihub data is ONLY permit to Korean!!!!!!! <br />
aihub์—์„œ ๋ฐ์ดํ„ฐ๋ฅผ ๋ฐ›์œผ์‹ค ๋ถ„์€ ํ•œ๊ตญ์ธ์ผ ๊ฒƒ์ด๋ฏ€๋กœ, ํ•œ๊ธ€๋กœ๋งŒ ์ž‘์„ฑํ•ฉ๋‹ˆ๋‹ค. <br />
์ •ํ™•ํžˆ๋Š” ์Œ์„ฑ์ „์‚ฌ๋ฅผ ์ฒ ์ž์ „์‚ฌ๋กœ ๋ฒˆ์—ญํ•˜๋Š” ํ˜•ํƒœ๋กœ ํ•™์Šต๋œ ๋ชจ๋ธ์ž…๋‹ˆ๋‹ค. (ETRI ์ „์‚ฌ๊ธฐ์ค€) <br />
- In case, ten million, some people use 10 million or some people use 10000000, so this model is crucial for training datasets <br />
์ฒœ๋งŒ์„ 1000๋งŒ ํ˜น์€ 10000000์œผ๋กœ ์“ธ ์ˆ˜๋„ ์žˆ๊ธฐ์—, Training Datasets์— ๋”ฐ๋ผ ๊ฒฐ๊ณผ๋Š” ์ƒ์ดํ•  ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค. <br />
- **์ˆ˜๊ด€ํ˜•์‚ฌ์™€ ์ˆ˜ ์˜์กด๋ช…์‚ฌ์˜ ๋„์–ด์“ฐ๊ธฐ์— ๋”ฐ๋ผ ๊ฒฐ๊ณผ๊ฐ€ ํ™•์—ฐํžˆ ๋‹ฌ๋ผ์งˆ ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค. (์‰ฐ์‚ด, ์‰ฐ ์‚ด -> ์‰ฐ์‚ด, 50์‚ด)** https://eretz2.tistory.com/34 <br />
์ผ๋‹จ์€ ๊ธฐ์ค€์„ ์žก๊ณ  ์น˜์šฐ์น˜๊ฒŒ ํ•™์Šต์‹œํ‚ค๊ธฐ์—” ์–ด๋–ป๊ฒŒ ์‚ฌ์šฉ๋ ์ง€ ๋ชฐ๋ผ, ํ•™์Šต ๋ฐ์ดํ„ฐ ๋ถ„ํฌ์— ๋งก๊ธฐ๋„๋ก ํ–ˆ์Šต๋‹ˆ๋‹ค. (์‰ฐ ์‚ด์ด ๋” ๋งŽ์„๊นŒ ์‰ฐ์‚ด์ด ๋” ๋งŽ์„๊นŒ!?)
- **Developed by:** Yoo SungHyun(https://github.com/YooSungHyun)
- **Language(s):** Korean
- **License:** apache-2.0
- **Parent Model:** See the [kobart-base-v2](https://huggingface.co/gogamza/kobart-base-v2) for more information about the pre-trained base model.
## Uses
Want see more detail follow this URL [KoGPT_num_converter](https://github.com/ddobokki/KoGPT_num_converter) <br /> and see `bart_inference.py` and `bart_train.py`
## Evaluation
Just using `evaluate-metric/bleu` and `evaluate-metric/rouge` in huggingface `evaluate` library <br />
[Training wanDB URL](https://wandb.ai/bart_tadev/BartForConditionalGeneration/runs/326xgytt?workspace=user-bart_tadev)
## How to Get Started With the Model
```python
from transformers.pipelines import Text2TextGenerationPipeline
from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
texts = ["๊ทธ๋Ÿฌ๊ฒŒ ๋ˆ„๊ฐ€ 6์‹œ๊นŒ์ง€ ์ˆ ์„ ๋งˆ์‹œ๋ž˜?"]
tokenizer = AutoTokenizer.from_pretrained("lIlBrother/ko-barTNumText")
model = AutoModelForSeq2SeqLM.from_pretrained("lIlBrother/ko-barTNumText")
seq2seqlm_pipeline = Text2TextGenerationPipeline(model=model, tokenizer=tokenizer)
kwargs = {
"min_length": 0,
"max_length": 1206,
"num_beams": 100,
"do_sample": False,
"num_beam_groups": 1,
}
pred = seq2seqlm_pipeline(texts, **kwargs)
print(pred)
# ๊ทธ๋Ÿฌ๊ฒŒ ๋ˆ„๊ฐ€ ์—ฌ์„ฏ ์‹œ๊นŒ์ง€ ์ˆ ์„ ๋งˆ์‹œ๋ž˜?
```