Downloads · 30 days
2
1% of all-time downloads
llama31-8b-ko-lora-gguf
shinyuuuu/llama31-8b-ko-lora-gguf
llama31-8b-ko-lora-gguf is a machine learning model from shinyuuuu. Use it for the machine learning task on the model card, and read the license before you ship it in a product. It is set up for transformers. The card lists the license as llama3.1.
Meta의 Llama 3.1 8B Instruct에 한글 데이터로 LoRA (SFT) 학습 후 병합한 모델입니다.
All-time downloads
183
Public
Repo size
4.9 GB
Likes
1
Public
Repo makeup
Click a slice to open those files.
.gguf4.9 GB · 100%
At a glance
- Library
- transformers
- License
- llama3.1
- Access
- Public
- Created
- Jul 17, 2026
- Updated
- Jul 17, 2026
- SHA
- 7f5518c2
Base models
Model card
From the Hugging Face model README
Llama 3.1 8B Instruct — Korean LoRA
Meta의 Llama 3.1 8B Instruct에 한글 데이터로 LoRA (SFT) 학습 후 병합한 모델입니다.
학습 정보
| 항목 | 내용 |
|---|---|
| Base Model | meta-llama/Llama-3.1-8B-Instruct |
| 학습 방식 | QLoRA (4bit) + SFT |
| 학습 도구 | unsloth + TRL SFTTrainer |
| GPU | NVIDIA A40 (48GB) |
| 학습 시간 | 약 25분 |
| LoRA rank | 32 |
| LoRA alpha | 32 |
| 학습률 | 2e-4 |
| Epoch | 1 |
| Batch size | 32 (effective 64) |
| Max seq length | 2048 |
데이터셋
| 데이터셋 | 샘플 수 | 용도 |
|---|---|---|
| beomi/KoAlpaca-v1.1a | 5,000 | 한글 Instruction |
| kyujinpy/KOpen-platypus | 3,000 | 추론/CoT |
| heegyu/PKU-SafeRLHF-ko | 1,000 | 안전 데이터 |
총 9,000건
사용법
from transformers import AutoModelForCausalLM, AutoTokenizer
model = AutoModelForCausalLM.from_pretrained("shinyuuuu/llama31-8b-ko-lora")
tokenizer = AutoTokenizer.from_pretrained("shinyuuuu/llama31-8b-ko-lora")
GGUF 버전 (LM Studio 등): shinyuuuu/llama31-8b-ko-lora-gguf
라이선스
Llama 3.1 Community License에 따릅니다.
- Library
- transformers
- License
- llama3.1
- Languages
- ko, en
- Created
- Jul 17, 2026
- Updated
- Jul 17, 2026