<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. -->
KoT5_news_summarization
- This model is a lcw99/t5-base-korean-text-summary finetuned on the daekeun-ml/naver-news-summarization-ko
Model description
<<20221021 Commit>>
프로젝트용으로 뉴스 요약 모델 특화된 모델을 만들기 위해 lcw99님의 t5-base-korean-text-summary 모델에 추가적으로 daekeun-ml님이 제공해주신 naver-news-summarization-ko 데이터셋으로 파인튜닝 했습니다.
현재 제가 가지고 있는 뉴스 데이터로 추가 학습 진행 예정입니다. 지속적으로 발전시켜 좋은 성능의 모델을 구현하겠습니다. 감사합니다.
실행환경
- Google Colab Pro
- CPU : Intel(R) Xeon(R) CPU @ 2.20GHz
- GPU : A100-SXM4-40GB
<pre><code>
Python Code
from transformers import AutoTokenizer from transformers import AutoModelForSeq2SeqLM
tokenizer = AutoTokenizer.from_pretrained("noahkim/KoT5_news_summarization") model = AutoModelForSeq2SeqLM.from_pretrained("noahkim/KoT5_news_summarization") </pre></code>
Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 2e-05
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 4
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 0.4513 | 1.0 | 2775 | 0.4067 |
| 0.42 | 2.0 | 5550 | 0.3933 |
| 0.395 | 3.0 | 8325 | 0.3864 |
| 0.3771 | 4.0 | 11100 | 0.3872 |
Framework versions
- Transformers 4.23.1
- Pytorch 1.12.1+cu113
- Datasets 2.6.1
- Tokenizers 0.13.1
