๐Ÿฆ… Pretrained BigBird Model for Korean (up to 4096 tokens)

Overview

Pretrained BigBird Model for Korean

What is BigBird โ€ข How to Use โ€ข Pretraining โ€ข Evaluation Result โ€ข Docs โ€ข Citation

ํ•œ๊ตญ์–ด | English

Apache 2.0 Issues linter DOI

What is BigBird?

BigBird: Transformers for Longer Sequences์—์„œ ์†Œ๊ฐœ๋œ sparse-attention ๊ธฐ๋ฐ˜์˜ ๋ชจ๋ธ๋กœ, ์ผ๋ฐ˜์ ์ธ BERT๋ณด๋‹ค ๋” ๊ธด sequence๋ฅผ ๋‹ค๋ฃฐ ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค.

๐Ÿฆ… Longer Sequence - ์ตœ๋Œ€ 512๊ฐœ์˜ token์„ ๋‹ค๋ฃฐ ์ˆ˜ ์žˆ๋Š” BERT์˜ 8๋ฐฐ์ธ ์ตœ๋Œ€ 4096๊ฐœ์˜ token์„ ๋‹ค๋ฃธ

โฑ๏ธ Computational Efficiency - Full attention์ด ์•„๋‹Œ Sparse Attention์„ ์ด์šฉํ•˜์—ฌ O(n2)์—์„œ O(n)์œผ๋กœ ๊ฐœ์„ 

How to Use

  • ๐Ÿค— Huggingface Hub์— ์—…๋กœ๋“œ๋œ ๋ชจ๋ธ์„ ๊ณง๋ฐ”๋กœ ์‚ฌ์šฉํ•  ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค:)
  • ์ผ๋ถ€ ์ด์Šˆ๊ฐ€ ํ•ด๊ฒฐ๋œ transformers>=4.11.0 ์‚ฌ์šฉ์„ ๊ถŒ์žฅํ•ฉ๋‹ˆ๋‹ค. (MRC ์ด์Šˆ ๊ด€๋ จ PR)
  • BigBirdTokenizer ๋Œ€์‹ ์— BertTokenizer ๋ฅผ ์‚ฌ์šฉํ•ด์•ผ ํ•ฉ๋‹ˆ๋‹ค. (AutoTokenizer ์‚ฌ์šฉ์‹œ BertTokenizer๊ฐ€ ๋กœ๋“œ๋ฉ๋‹ˆ๋‹ค.)
  • ์ž์„ธํ•œ ์‚ฌ์šฉ๋ฒ•์€ BigBird Tranformers documentation์„ ์ฐธ๊ณ ํ•ด์ฃผ์„ธ์š”.
from transformers import AutoModel, AutoTokenizer

model = AutoModel.from_pretrained("monologg/kobigbird-bert-base")  # BigBirdModel
tokenizer = AutoTokenizer.from_pretrained("monologg/kobigbird-bert-base")  # BertTokenizer

Pretraining

์ž์„ธํ•œ ๋‚ด์šฉ์€ [Pretraining BigBird] ์ฐธ๊ณ 

Hardware Max len LR Batch Train Step Warmup Step
KoBigBird-BERT-Base TPU v3-8 4096 1e-4 32 2M 20k
  • ๋ชจ๋‘์˜ ๋ง๋ญ‰์น˜, ํ•œ๊ตญ์–ด ์œ„ํ‚ค, Common Crawl, ๋‰ด์Šค ๋ฐ์ดํ„ฐ ๋“ฑ ๋‹ค์–‘ํ•œ ๋ฐ์ดํ„ฐ๋กœ ํ•™์Šต
  • ITC (Internal Transformer Construction) ๋ชจ๋ธ๋กœ ํ•™์Šต (ITC vs ETC)

Evaluation Result

1. Short Sequence (<=512)

์ž์„ธํ•œ ๋‚ด์šฉ์€ [Finetune on Short Sequence Dataset] ์ฐธ๊ณ 

NSMC
(acc)
KLUE-NLI
(acc)
KLUE-STS
(pearsonr)
Korquad 1.0
(em/f1)
KLUE MRC
(em/rouge-w)
KoELECTRA-Base-v3 91.13 86.87 93.14 85.66 / 93.94 59.54 / 65.64
KLUE-RoBERTa-Base 91.16 86.30 92.91 85.35 / 94.53 69.56 / 74.64
KoBigBird-BERT-Base 91.18 87.17 92.61 87.08 / 94.71 70.33 / 75.34

2. Long Sequence (>=1024)

์ž์„ธํ•œ ๋‚ด์šฉ์€ [Finetune on Long Sequence Dataset] ์ฐธ๊ณ 

TyDi QA
(em/f1)
Korquad 2.1
(em/f1)
Fake News
(f1)
Modu Sentiment
(f1-macro)
KLUE-RoBERTa-Base 76.80 / 78.58 55.44 / 73.02 95.20 42.61
KoBigBird-BERT-Base 79.13 / 81.30 67.77 / 82.03 98.85 45.42

Docs

Citation

KoBigBird๋ฅผ ์‚ฌ์šฉํ•˜์‹ ๋‹ค๋ฉด ์•„๋ž˜์™€ ๊ฐ™์ด ์ธ์šฉํ•ด์ฃผ์„ธ์š”.

@software{jangwon_park_2021_5654154,
  author       = {Jangwon Park and Donggyu Kim},
  title        = {KoBigBird: Pretrained BigBird Model for Korean},
  month        = nov,
  year         = 2021,
  publisher    = {Zenodo},
  version      = {1.0.0},
  doi          = {10.5281/zenodo.5654154},
  url          = {https://doi.org/10.5281/zenodo.5654154}
}

Contributors

Jangwon Park and Donggyu Kim

Acknowledgements

KoBigBird๋Š” Tensorflow Research Cloud (TFRC) ํ”„๋กœ๊ทธ๋žจ์˜ Cloud TPU ์ง€์›์œผ๋กœ ์ œ์ž‘๋˜์—ˆ์Šต๋‹ˆ๋‹ค.

๋˜ํ•œ ๋ฉ‹์ง„ ๋กœ๊ณ ๋ฅผ ์ œ๊ณตํ•ด์ฃผ์‹  Seyun Ahn๋‹˜๊ป˜ ๊ฐ์‚ฌ๋ฅผ ์ „ํ•ฉ๋‹ˆ๋‹ค.

You might also like...
KakaoBrain KoGPT (Korean Generative Pre-trained Transformer)

KoGPT KoGPT (Korean Generative Pre-trained Transformer) https://github.com/kakaobrain/kogpt https://huggingface.co/kakaobrain/kogpt Model Descriptions

Generating Korean Slogans with phonetic and structural repetition
Generating Korean Slogans with phonetic and structural repetition

LexPOS_ko Generating Korean Slogans with phonetic and structural repetition Generating Slogans with Linguistic Features LexPOS is a sequence-to-sequen

Korean extractive summarization. 2021 AI ํ…์ŠคํŠธ ์š”์•ฝ ์˜จ๋ผ์ธ ํ•ด์ปคํ†ค ํ™”์„ฑ๊ฐˆ๋„๋‹ˆ๊นŒํŒ€ ์ฝ”๋“œ
Korean extractive summarization. 2021 AI ํ…์ŠคํŠธ ์š”์•ฝ ์˜จ๋ผ์ธ ํ•ด์ปคํ†ค ํ™”์„ฑ๊ฐˆ๋„๋‹ˆ๊นŒํŒ€ ์ฝ”๋“œ

korean extractive summarization 2021 AI ํ…์ŠคํŠธ ์š”์•ฝ ์˜จ๋ผ์ธ ํ•ด์ปคํ†ค ํ™”์„ฑ๊ฐˆ๋„๋‹ˆ๊นŒํŒ€ ์ฝ”๋“œ Leaderboard Notice Text Summarization with Pretrained Encoders์— ๋‚˜์˜ค๋Š” bertsumext๋ชจ๋ธ(ext

Training code for Korean multi-class sentiment analysis

KoSentimentAnalysis Bert implementation for the Korean multi-class sentiment analysis ์™œ ํ•œ๊ตญ์–ด ๊ฐ์ • ๋‹ค์ค‘๋ถ„๋ฅ˜ ๋ชจ๋ธ์€ ๊ฑฐ์˜ ์—†๋Š” ๊ฒƒ์ผ๊นŒ?์—์„œ ์‹œ์ž‘๋œ ํ”„๋กœ์ ํŠธ Environment: Pytorch, Da

Korean Sentence Embedding Repository

Korean-Sentence-Embedding ๐Ÿญ Korean sentence embedding repository. You can download the pre-trained models and inference right away, also it provides

ProteinBERT is a universal protein language model pretrained on ~106M proteins from the UniRef90 dataset.

ProteinBERT is a universal protein language model pretrained on ~106M proteins from the UniRef90 dataset. Through its Python API, the pretrained model can be fine-tuned on any protein-related task in a matter of minutes. Based on our experiments with a wide range of benchmarks, ProteinBERT usually achieves state-of-the-art performance. ProteinBERT is built on TenforFlow/Keras.

IndoBERTweet is the first large-scale pretrained model for Indonesian Twitter. Published at EMNLP 2021 (main conference)

IndoBERTweet ๐Ÿฆ ๐Ÿ‡ฎ๐Ÿ‡ฉ 1. Paper Fajri Koto, Jey Han Lau, and Timothy Baldwin. IndoBERTweet: A Pretrained Language Model for Indonesian Twitter with Effe

BMInf (Big Model Inference) is a low-resource inference package for large-scale pretrained language models (PLMs).
BMInf (Big Model Inference) is a low-resource inference package for large-scale pretrained language models (PLMs).

BMInf (Big Model Inference) is a low-resource inference package for large-scale pretrained language models (PLMs).

Crie tokens de autenticaรงรฃo รญntegros e seguros com UToken.

UToken - Tokens seguros. UToken (ou Unhandleable Token) รฉ uma bilioteca criada para ser utilizada na geraรงรฃo de tokens seguros e รญntegros, ou seja, nรฃ

Comments
  • Pretraining Epoch ์งˆ๋ฌธ

    Pretraining Epoch ์งˆ๋ฌธ

    Checklist

    • [x] I've searched the project's issues

    โ“ Question

    ์•ˆ๋…•ํ•˜์„ธ์š” ์ €๋Š” ํ˜„์žฌ ์นœ๊ตฌ๋“ค๊ณผ ํ•จ๊ป˜ 4096 ํ† ํฐ์„ ์ž…๋ ฅ๋ฐ›์•„ ์š”์•ฝ ํƒœ์Šคํฌ๋ฅผ ์ˆ˜ํ–‰ํ•  ์ˆ˜ ์žˆ๋Š” ๋ชจ๋ธ์„ ๋งŒ๋“ค๊ณ  ์žˆ์Šต๋‹ˆ๋‹ค. ์ฒ˜์Œ์—” ๋น…๋ฒ„๋“œ + ๋ฒ„ํŠธ ์กฐํ•ฉ์œผ๋กœ ํ•ด๋ณด๋ ค๊ณ  ํ–ˆ๋Š”๋ฐ, ์ด๋ฏธ monologg ๋‹˜๊ป˜์„œ ๋งŒ๋“ค์–ด์ฃผ์…จ๋”๋ผ๊ตฌ์š” ใ…Žใ…Ž ๊ทธ๋ž˜์„œ ๋กฑํฌ๋จธ + ๋ฐ”ํŠธ + ํŽ˜๊ฐ€์ˆ˜์Šค ์กฐํ•ฉ์œผ๋กœ ํ•™์Šต์„ ์ง„ํ–‰ํ•˜๋ ค ํ•˜๊ณ  ์žˆ์Šต๋‹ˆ๋‹ค. pretrained๋œ KoBart๋ฅผ ๊ธฐ๋ฐ˜์œผ๋กœ ์–ดํ…์…˜์„ ๋กฑํฌ๋จธ๋กœ ๋ฐ”๊พผ ํ›„, ํŽ˜๊ฐ€์ˆ˜์Šค task๋ฅผ ์ˆ˜ํ–‰ํ•˜๋Š” ๊ตฌ์กฐ๋กœ ๋˜์–ด ์žˆ์Šต๋‹ˆ๋‹ค.

    ํ˜„์žฌ 13GB์˜ ๋ฐ์ดํ„ฐ๋ฅผ ๋ชจ์•„์„œ ์ „์ฒ˜๋ฆฌ์™€ ๋ฐ์ดํ„ฐ๋กœ๋” ์ž‘์„ฑ, ๋ชจ๋ธ ์ฝ”๋“œ๊นŒ์ง€๋Š” ์™„๋ฃŒํ•œ ์ƒํƒœ์ž…๋‹ˆ๋‹ค. ์ด๋ฒˆ ์ฃผ ๋‚ด๋กœ ํ•™์Šต์„ ์ง„ํ–‰ํ•˜๋ ค ํ•˜๊ณ  ์žˆ์Šต๋‹ˆ๋‹ค.

    ์ €ํฌ๊ฐ€ ๊ฐ€์ง„ GPU๋กœ๋Š” ๋Œ€๋žต ์ดํ‹€์ด๋ฉด 1 ์—ํฌํฌ๋ฅผ ๋Œ ์ˆ˜ ์žˆ์„ ๊ฒƒ ๊ฐ™์€๋ฐ, monologg๋‹˜๊ป˜์„œ๋Š” KoBirBird ๋ชจ๋ธ ๊ฐœ๋ฐœ ์‹œ ์—ํฌํฌ๋ฅผ ์–ผ๋งˆ๋‚˜ ๋„์…จ๋Š”์ง€ ์—ฌ์ญค๋ณด๊ณ  ์‹ถ์Šต๋‹ˆ๋‹ค.

    ์•„๋ฌด๋ž˜๋„ pretrained ๋œ ๋ชจ๋ธ์„ ๊ฐ€์ ธ๋‹ค ์“ฐ๋‹ค๋ณด๋‹ˆ ์—ํฌํฌ๋ฅผ ๋งŽ์ด ๋Œ ํ•„์š”๋Š” ์—†์„ ๊ฒƒ ๊ฐ™์€๋ฐ, ๊ธฐ์ค€์ ์œผ๋กœ ์‚ผ๊ณ  ์‹ถ์–ด์„œ์š”!

    ๋ง์ด ๊ธธ์–ด์กŒ๋Š”๋ฐ ์š”์•ฝํ•˜์ž๋ฉด, KoBirBird ํ•™์Šต ์‹œ ์—ํฌํฌ๋ฅผ ์–ผ๋งˆ๋‚˜ ์ฃผ์…จ๋Š”์ง€ ๊ถ๊ธˆํ•ฉ๋‹ˆ๋‹ค. ๋˜ํ•œ, ๊ทธ ๊ธฐ์ค€์€ ๋ฌด์—‡์œผ๋กœ ์‚ผ์œผ์…จ๋Š”์ง€๋„ ๊ถ๊ธˆํ•ฉ๋‹ˆ๋‹ค.

    question 
    opened by KimJaehee0725 2
  • Specific information about this model.

    Specific information about this model.

    Checklist

    • [ x ] I've searched the project's issues

    โ“ Question

    • You mentioned "๋ชจ๋‘์˜ ๋ง๋ญ‰์น˜, ํ•œ๊ตญ์–ด ์œ„ํ‚ค, Common Crawl, ๋‰ด์Šค ๋ฐ์ดํ„ฐ ๋“ฑ ๋‹ค์–‘ํ•œ ๋ฐ์ดํ„ฐ๋กœ ํ•™์Šต" and I want to know the size of total corpus for pre-training.

    • Also I want to know the vocab size of this model.

    ๐Ÿ“Ž Additional context

    question 
    opened by midannii 2
  • Fix some minors

    Fix some minors

    Description

    ์ฝ”๋“œ์™€ ์ฃผ์„ ๋“ฑ์„ ์ฝ๋‹ค๊ฐ€ ๋ณด์ธ ์ž‘์€ ์˜คํƒ€ ๋“ฑ์„ ์ˆ˜์ •ํ–ˆ์Šต๋‹ˆ๋‹ค

    ๋‹ค์–‘ํ•œ ๋…ธํ•˜์šฐ๋ฅผ ์•„๋‚Œ์—†์ด ๊ณต์œ ํ•ด์ฃผ์‹  @monologg , @donggyukimc ์—๊ฒŒ ๊ฐ์‚ฌ์˜ ๋ง์”€๋“œ๋ฆฝ๋‹ˆ๋‹ค.

    ์ดํ›„์—๋Š” GPU ํ™˜๊ฒฝ์—์„œ finetuning์„ ํ…Œ์ŠคํŠธํ•ด ๋ณผ ์˜ˆ์ •์ž…๋‹ˆ๋‹ค ๊ณ ๋ง™์Šต๋‹ˆ๋‹ค.

    Related Issue

    chore 
    opened by sackoh 0
Releases(v1.0.0)
Utilize Korean BERT model in sentence-transformers library

ko-sentence-transformers ์ด ํ”„๋กœ์ ํŠธ๋Š” KoBERT ๋ชจ๋ธ์„ sentence-transformers ์—์„œ ๋ณด๋‹ค ์‰ฝ๊ฒŒ ์‚ฌ์šฉํ•˜๊ธฐ ์œ„ํ•ด ๋งŒ๋“ค์–ด์กŒ์Šต๋‹ˆ๋‹ค. Ko-Sentence-BERT-SKTBERT ํ”„๋กœ์ ํŠธ์—์„œ๋Š” KoBERT ๋ชจ๋ธ์„ sentence-trans

Junghyun 40 Dec 20, 2022
A collection of Korean Text Datasets ready to use using Tensorflow-Datasets.

tfds-korean A collection of Korean Text Datasets ready to use using Tensorflow-Datasets. TensorFlow-Datasets๋ฅผ ์ด์šฉํ•œ ํ•œ๊ตญ์–ด/ํ•œ๊ธ€ ๋ฐ์ดํ„ฐ์…‹ ๋ชจ์Œ์ž…๋‹ˆ๋‹ค. Dataset Catalog |

Jeong Ukjae 20 Jul 11, 2022
Baseline code for Korean open domain question answering(ODQA)

Open-Domain Question Answering(ODQA)๋Š” ๋‹ค์–‘ํ•œ ์ฃผ์ œ์— ๋Œ€ํ•œ ๋ฌธ์„œ ์ง‘ํ•ฉ์œผ๋กœ๋ถ€ํ„ฐ ์ž์—ฐ์–ด ์งˆ์˜์— ๋Œ€ํ•œ ๋‹ต๋ณ€์„ ์ฐพ์•„์˜ค๋Š” task์ž…๋‹ˆ๋‹ค. ์ด๋•Œ ์‚ฌ์šฉ์ž ์งˆ์˜์— ๋‹ต๋ณ€ํ•˜๊ธฐ ์œ„ํ•ด ์ฃผ์–ด์ง€๋Š” ์ง€๋ฌธ์ด ๋”ฐ๋กœ ์กด์žฌํ•˜์ง€ ์•Š์Šต๋‹ˆ๋‹ค. ๋”ฐ๋ผ์„œ ์‚ฌ์ „์— ๊ตฌ์ถ•๋˜์–ด์žˆ๋Š” Knowl

VUMBLEB 69 Nov 4, 2022
A BERT-based reverse-dictionary of Korean proverbs

Wisdomify A BERT-based reverse-dictionary of Korean proverbs. ๊น€์œ ๋นˆ : ๋ชจ๋ธ๋ง / ๋ฐ์ดํ„ฐ ์ˆ˜์ง‘ / ํ”„๋กœ์ ํŠธ ์„ค๊ณ„ / back-end ๊น€์ข…์œค : ๋ฐ์ดํ„ฐ ์ˆ˜์ง‘ / ํ”„๋กœ์ ํŠธ ์„ค๊ณ„ / front-end Quick Start C

Eu-Bin KIM 94 Dec 8, 2022
Official implementation of MLP Singer: Towards Rapid Parallel Korean Singing Voice Synthesis

MLP Singer Official implementation of MLP Singer: Towards Rapid Parallel Korean Singing Voice Synthesis. Audio samples are available on our demo page.

Neosapience 103 Dec 23, 2022
KLUE-baseline contains the baseline code for the Korean Language Understanding Evaluation (KLUE) benchmark.

KLUE Baseline Korean(ํ•œ๊ตญ์–ด) KLUE-baseline contains the baseline code for the Korean Language Understanding Evaluation (KLUE) benchmark. See our paper fo

null 74 Dec 13, 2022
Korean Simple Contrastive Learning of Sentence Embeddings using SKT KoBERT and kakaobrain KorNLU dataset

KoSimCSE Korean Simple Contrastive Learning of Sentence Embeddings implementation using pytorch SimCSE Installation git clone https://github.com/BM-K/

null 34 Nov 24, 2022
A BERT-based reverse dictionary of Korean proverbs

Wisdomify A BERT-based reverse-dictionary of Korean proverbs. ๊น€์œ ๋นˆ : ๋ชจ๋ธ๋ง / ๋ฐ์ดํ„ฐ ์ˆ˜์ง‘ / ํ”„๋กœ์ ํŠธ ์„ค๊ณ„ / back-end ๊น€์ข…์œค : ๋ฐ์ดํ„ฐ ์ˆ˜์ง‘ / ํ”„๋กœ์ ํŠธ ์„ค๊ณ„ / front-end / back-end ์ž„์šฉ

null 94 Dec 8, 2022
Korean stereoypte detector with TUNiB-Electra and K-StereoSet

Korean Stereotype Detector Korean stereotype sentence classifier using K-StereoSet with TUNiB-Electra Web demo you can test this model easily in demo

Sae_Chan_Oh 11 Feb 18, 2022
Transformer Based Korean Sentence Spacing Corrector

TKOrrector Transformer Based Korean Sentence Spacing Corrector License Summary This solution is made available under Apache 2 license. See the LICENSE

Paul Hyung Yuel Kim 3 Apr 18, 2022