# Beomi/KoAlpaca

KoAlpaca: 한국어 명령어를 이해하는 오픈소스 언어모델 (KoAlpaca: An open-source language model to understand Korean instructions)

Repository: https://github.com/Beomi/KoAlpaca
Canonical: https://ross.abutalabs.com/products/koalpaca
Language: Jupyter Notebook
License: Apache-2.0
License Family: permissive
Topics: alpaca, koalpaca, korean-nlp, llama, chatkoalpaca, polyglot-ko
Last push: 2024-10-25T07:59:42+00:00

## Health v2 (maintenance only)
Score: 30/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 0, release rhythm 35, longevity 90
- inputs: {"age_days": 1264, "days_push": 677, "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 1573, forks 224 (observed 2026-08-28T04:05:05.628482+00:00)

## What it is
KoAlpaca is an open-source Korean instruction-following language model project with training code, datasets, and fine-tuned model weights based on LLaMA and Polyglot-ko. It includes Colab tutorials for QLoRA fine-tuning and inference on consumer GPUs.

## Use cases
- fine-tune a Korean instruction-following LLM
- train an Alpaca-style model on Korean data
- run QLoRA fine-tuning on a free Colab T4 GPU
- get a Korean chatbot language model
- fine-tune Polyglot-ko on custom instructions
- learn LLM fine-tuning with Korean datasets

## When to choose
- you need a Korean-language instruction-tuned open model
- you want reproducible fine-tuning examples for LLaMA or Polyglot-ko on limited GPU hardware
- you need the KoAlpaca v1.1 instruction dataset

## When to avoid
- you need a hosted chat demo (the web demo was shut down in May 2024)
- you need a production-ready multilingual assistant rather than a research model
- you need non-Korean language support

## Facets
- artifact type: learning-resource
- maturity: maintenance
- function: llm-training, machine-learning, nlp
- domain: large-language-models, machine-learning
- platform: python
- tags: korean, alpaca, llama, instruction-tuning, finetuning, qlora, polyglot-ko, dataset, natural-language-processing, gpu

## Member repositories
- Beomi/KoAlpaca (main) score 30

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:05:05.628482+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 2026-08-30T03:57:56.044516+00:00, confidence not recorded.
  - readme: https://github.com/Beomi/KoAlpaca (fetched 2026-08-28T04:05:05.628482+00:00, sha a9de58822854)
- Data as of 2026-08-30T08:39:29.467469+00:00.