---
license: cc-by-nc-4.0
datasets:
- MarkrAI/KOpen-HQ-Hermes-2.5-60K
language:
- ko
base_model:
- meta-llama/Meta-Llama-3.1-8B-Instruct
pipeline_tag: text-generation
library_name: adapter-transformers
---
A Creative Commons NonCommercial license (CC NC, CC BY-NC or NC license) is a Creative Commons license which a copyright holder can apply to their media to give public permission for anyone to reuse that media only for noncommercial activities. Creative Commons is an organization which develops a variety of public copyright licenses, and the "noncommercial" licenses are a subset of these. Unlike the CC0, CC BY, and CC BY-SA licenses, the CC BY-NC license is considered non-free.

unsloth를 사용하여 meta-llama/Meta-Llama-3.1-8B-Instruct 모델에 LORA 파인튜닝을 완료했습니다.

MarkrAI/KOpen-HQ-Hermes-2.5-60k 데이터를 학습시켰습니다.

## How to use

```python
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("NAPS-ai/naps-llama-3.1-8b-instruct-v0.4")
model = AutoModelForCausalLM.from_pretrained("NAPS-ai/naps-llama-3.1-8b-instruct-v0.4")
```

## Chatbot


![image/png](https://cdn-uploads.huggingface.co/production/uploads/6699b80354725cd6e0ae8e19/1J506GxR0eT6XnKsGVbye.png)

```python
from transformers import AutoTokenizer, AutoModelForCausalLM
import transformers
import torch

tokenizer = AutoTokenizer.from_pretrained("NAPS-ai/naps-llama-3.1-8b-instruct-v0.4")
model = AutoModelForCausalLM.from_pretrained("NAPS-ai/naps-llama-3.1-8b-instruct-v0.4")




pipeline = transformers.pipeline(
    "text-generation",
    model=model,
    tokenizer=tokenizer,
    model_kwargs={"torch_dtype": torch.bfloat16},
    device=0,
)

def answering(question):
    messages = [
    {"role": "system", "content": "당신은 항상 친절하게 대답하는 안내원입니다."},
    {"role": "user", "content": question},
    ]
    outputs = pipeline(
    messages,
    max_new_tokens=1024,
    pad_token_id = pipeline.tokenizer.eos_token_id
    )
    return outputs[0]["generated_text"][2]['content']




while True:
    question = input("질문을 입력하세요 : ")
    if question == "종료":
        print("프로그램 종료")
        break
    answer = answering(question)
    print(f"AI의 답변: {answer}")

```


Contact : cgh@tnap.co.kr