# SKTBrain/KoBERT

Korean BERT pre-trained cased (KoBERT)

Repository: https://github.com/SKTBrain/KoBERT
Canonical: https://ross.abutalabs.com/products/kobert
Language: Python
License: Apache-2.0
License Family: permissive
Topics: korean-nlp, language-model, bert, nlp, pytorch, transformers
Last push: 2025-06-14T19:06:29+00:00

## Health v2 (maintenance only)
Score: 44/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 26, release rhythm 35, longevity 100
- inputs: {"age_days": 2536, "days_push": 445, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1416, forks 377 (observed 2026-08-28T04:04:39.953124+00:00)

## What it is
KoBERT is a Korean BERT language model pre-trained on Korean Wikipedia by SK Telecom, distributed as a Python package with PyTorch, ONNX, and MXNet/Gluon interfaces plus a SentencePiece tokenizer. It also provides subtask models such as Naver sentiment analysis, Korean NER with CRF, and Korean Sentence BERT.

## Use cases
- fine-tune BERT on Korean text
- Korean sentiment analysis
- Korean named entity recognition
- generate Korean sentence embeddings
- use a Korean language model in PyTorch or ONNX

## When to choose
- you need a compact, well-tested Korean BERT encoder for classification, NER, or embeddings
- you work in PyTorch or Hugging Face transformers and want Korean pretraining

## When to avoid
- you need a multilingual or newest-generation Korean model with larger vocabulary
- you need active feature development rather than a maintained pretrained model

## Facets
- artifact type: library
- maturity: maintenance
- function: machine-learning, nlp, transformers
- domain: machine-learning, deep-learning
- platform: python
- tags: korean-nlp, bert, language-model, pytorch, sentencepiece, pretrained-model, natural-language-processing, gpu

## Member repositories
- SKTBrain/KoBERT (main) score 44

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:04:39.953124+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T04:38:04.455587+00:00, confidence not recorded.
  - readme: https://github.com/SKTBrain/KoBERT (fetched 2026-08-28T04:04:39.953124+00:00, sha bb156b115aef)
- Data as of 2026-08-30T08:39:29.467469+00:00.