Bert base model for Korean

Update

  • Update at 2021.11.17 : Add Native Support for BERT Tokenizer (works with AutoTokenizer, pipeline)

  • 70GB Korean text dataset and 42000 lower-cased subwords are used
  • Check the model performance and other language models for Korean in github
from transformers import pipeline

pipe = pipeline('text-generation', model='beomi/kykim-gpt3-kor-small_based_on_gpt2')
print(pipe("μ•ˆλ…•ν•˜μ„Έμš”! μ˜€λŠ˜μ€"))
# [{'generated_text': 'μ•ˆλ…•ν•˜μ„Έμš”! μ˜€λŠ˜μ€ μ œκ°€ μš”μ¦˜ μ‚¬μš©ν•˜κ³  μžˆλŠ” ν΄λ Œμ§•μ›Œν„°λ₯Ό μ†Œκ°œν•΄λ“œλ¦¬λ €κ³  ν•΄μš”! λ°”λ‘œ 이 μ œν’ˆ!! λ°”λ‘œ 이'}]
Downloads last month
127
Inference Examples
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social visibility and check back later, or deploy to Inference Endpoints (dedicated) instead.