https://github.com/sooftware/korean-plm
List of Korean pre-trained language models.
https://github.com/sooftware/korean-plm
Last synced: 2 months ago
JSON representation
List of Korean pre-trained language models.
- Host: GitHub
- URL: https://github.com/sooftware/korean-plm
- Owner: sooftware
- License: mit
- Created: 2022-07-12T16:43:27.000Z (almost 3 years ago)
- Default Branch: main
- Last Pushed: 2023-08-31T15:02:37.000Z (over 1 year ago)
- Last Synced: 2025-01-24T06:13:24.090Z (4 months ago)
- Homepage:
- Size: 26.4 KB
- Stars: 188
- Watchers: 5
- Forks: 17
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# Korean-PLM (Korean Pre-trained Language Models)
공개된 한국어 사전학습 모델을 기록합니다. 크게 아래 3개의 모델 계열로 구분했으며, 모델 사이즈는 정확하지 않을 수 있습니다.
- Encoder Model (BERT 계열)
- Decoder Model (GPT 계열)
- Encoder-Decoder Model (Seq2seq 계열)
## Encoder Model (BERT 계열)| Model Name | Size | Link |
| :------------: | :------------: | :------------: |
| LASSL BERT | Small (10M) | [link](https://huggingface.co/lassl/bert-ko-small) |
| LASSL RoBERTa | Small (10M) | [link](https://huggingface.co/lassl/roberta-ko-small) |
| TUNiB ELECTRA (Ko) | Small (10M) | [link](https://huggingface.co/tunib/electra-ko-small) |
| TUNiB ELECTRA (Ko-En) | Small (10M) | [link](https://huggingface.co/tunib/electra-ko-en-small) |
| KoELECTRA | Small (10M) | [link](https://huggingface.co/monologg/koelectra-small-v3-discriminator) |
| KcELECTRA | Small (10M) | [link](https://huggingface.co/beomi/kcelectra-v2022-dev) |
| Soongsil-BERT | Small (10M) | [link](https://huggingface.co/jason9693/soongsil-bert-small) |
| Ko-CHAR-ELECTRA | Small (10M) | [link](https://huggingface.co/monologg/kocharelectra-small-discriminator) |
| Distill-KoBERT | 30M | [link](https://huggingface.co/monologg/kocharelectra-base-discriminator) |
| KoBERT | Base (110M) | [link](https://huggingface.co/skt/kobert-base-v1) |
| KorBERT | Base (110M) | [link](https://aiopen.etri.re.kr/service_dataset.php) |
| HanBERT | Base (110M) | [link](https://github.com/monologg/HanBert-Transformers) |
| KcBERT | Base (110M) | [link](https://huggingface.co/beomi/kcbert-base) |
| Soongsil-BERT | Base (110M) | [link](https://huggingface.co/jason9693/soongsil-bert-base) |
| KLUE-RoBERTa | Base (110M) | [link](https://huggingface.co/klue/roberta-base) |
| KoELECTRA | Base (110M) | [link](https://huggingface.co/monologg/koelectra-base-v3-discriminator) |
| KcELECTRA | Base (110M) | [link](https://huggingface.co/beomi/KcELECTRA-base) |
| TUNiB-ELECTRA (Ko) | Base (110M) | [link](https://huggingface.co/tunib/electra-ko-base) |
| TUNiB-ELECTRA (Ko-En) | Base (110M) | [link](https://huggingface.co/tunib/electra-ko-en-base) |
| LMKor Albert | Base (110M) | [link](https://huggingface.co/kykim/albert-kor-base) |
| LMKor Bert | Base (110M) | [link](https://huggingface.co/kykim/bert-kor-base) |
| LMKor Funnel | Base (110M) | [link](https://huggingface.co/kykim/funnel-kor-base) |
| LMKor ELECTRA | Base (110M) | [link](https://huggingface.co/kykim/electra-kor-base) |
| KalBERT | Base (110M) | [link](https://github.com/MrBananaHuman/KalBert) |
| LASSL BERT | Base (110M) | [link](https://huggingface.co/lassl/bert-ko-base) |
| KoBigBird | Base (110M) | [link](https://huggingface.co/monologg/kobigbird-bert-base) |
| Ko-CHAR-BERT | Base (110M) | [link](https://github.com/MrBananaHuman/KoreanCharacterBert) |
| Ko-CHAR-ELECTRA | Base (110M) | [link](https://huggingface.co/monologg/kocharelectra-base-discriminator) |
| BrainSBERT | Base (110M) | [link](https://huggingface.co/hyunwoongko/brainsbert-base) |
| KPFBERT | Base (110M) | [link](https://huggingface.co/jinmang2/kpfbert) |
| KcBERT | Large (340M) | [link](https://huggingface.co/beomi/kcbert-large) |
| KLUE-RoBERTa | Large (340M) | [link](https://huggingface.co/klue/roberta-large) |## Decoder Model (GPT 계열)
| Model Name | Size | Link |
| :------------: | :------------: | :------------: |
| LASSL GPT2 | Small (20M) | [link](https://huggingface.co/lassl/gpt2-ko-small) |
| LMKor KoGPT2 | Base (110M) | [link](https://huggingface.co/kykim/gpt3-kor-small_based_on_gpt2) |
| SKT KoGPT2 | Base (110M) | [link](https://huggingface.co/skt/kogpt2-base-v2) |
| SKT KoGPT Trinity | 1.2B | [link](https://huggingface.co/skt/ko-gpt-trinity-1.2B-v0.5) |
| KakaoBrain KoGPT | 6B | [link](https://huggingface.co/kakaobrain/kogpt) |
| EleutherAI Polyglot-Ko | 1.3B | [link](https://huggingface.co/EleutherAI/polyglot-ko-1.3b) |
| EleutherAI Polyglot-Ko | 3.8B | [link](EleutherAI/polyglot-ko-3.8b) |
| EleutherAI Polyglot-Ko | 5.8B | [link](https://huggingface.co/EleutherAI/polyglot-ko-5.8b) |
| EleutherAI Polyglot-Ko | 12.8B | [link](https://huggingface.co/EleutherAI/polyglot-ko-12.8b) |
| KoAlpaca (Polyglot-Ko) | 5.8B | [link](https://huggingface.co/beomi/KoAlpaca-Polyglot-5.8B) |
| KoAlpaca (Polyglot-Ko) | 12.8B | [link](https://huggingface.co/beomi/KoAlpaca-Polyglot-12.8B) |## Encoder-Decoder Model (Seq2seq 계열)
| Model Name | Size | Link |
| :------------: | :------------: | :------------: |
| Cosmoquester BART | Mini (13M) | [link](https://huggingface.co/cosmoquester/bart-ko-mini) |
| KE-T5 (Ko-En) | Small (60M) | [link](https://huggingface.co/KETI-AIR/ke-t5-small) |
| KE-T5 (Ko) | Small (60M) | [link](https://huggingface.co/KETI-AIR/ke-t5-small-ko) |
| LMKor T5 | Small (60M) | [link](https://huggingface.co/kykim/t5-kor-small) |
| Cosmoquester BART | Small (40M) | [link](https://huggingface.co/cosmoquester/bart-ko-small) |
| LMKor BERT Shared | Base (130M) | [link](https://huggingface.co/kykim/bertshared-kor-base) |
| SKT KoBART | Base (130M) | [link](https://huggingface.co/gogamza/kobart-base-v2) |
| Cosmoquester BART | Base (130M) | [link](https://huggingface.co/gogamza/kobart-base-v2) |
| Kolang-T5 | Base (220M) | [link](https://huggingface.co/digit82/kolang-t5-base) |
| KcT5 | Base (220M) | [link](https://huggingface.co/beomi/KcT5-dev) |
| KE-T5 (Ko-En) | Base (220M) | [link](https://huggingface.co/KETI-AIR/ke-t5-base) |
| KE-T5 (Ko) | Base (220M) | [link](https://huggingface.co/KETI-AIR/ke-t5-basko) |
| Asian BART (Ko) | Base (360M) | [link](hyunwoongko/asian-bart-ko) |
| pko-t5 (Ko) | Base (250M) | [link](https://huggingface.co/paust/pko-t5-base) |
| Wisenut KoT5 (Ko) | Base (220M) | [link](https://github.com/wisenut-research/KoT5#usage) |
| mBART | Large (610M) | [link](https://huggingface.co/facebook/mbart-large-50) |
| KE-T5 (Ko-En) | Large (770M) | [link](https://huggingface.co/KETI-AIR/ke-t5-large) |
| KE-T5 (Ko) | Large (770M) | [link](https://huggingface.co/KETI-AIR/ke-t5-large-ko) |
| pko-t5 (Ko) | Large (800M) | [link](https://huggingface.co/paust/pko-t5-large) |