README.md

---
base_model: beomi/Llama-3-Open-Ko-8B
language:
- en
license: apache-2.0
tags:
- transformers
- unsloth
- llama
- trl
---

- **Developed by:** lwef
- **License:** apache-2.0
- **Finetuned from model :** beomi/Llama-3-Open-Ko-8B

# korean dialogue summary fine-tuned model
# how to use
```python
prompt_template = '''
아래 대화를 요약해 주세요. 대화 형식은 '#대화 참여자#: 대화 내용'입니다.
### 대화 >>>{dialogue}

### 요약 >>>'''

if True:
    from unsloth import FastLanguageModel
    model, tokenizer = FastLanguageModel.from_pretrained(
        model_name = "lwef/llm-bench-upload-1", # YOUR MODEL YOU USED FOR TRAINING
        max_seq_length = 2048,
        dtype = None,
        load_in_4bit = True,
    )
    FastLanguageModel.for_inference(model) # Enable native 2x faster inference
dialogue = '''#P01#: 아 행삶 과제 너무 어려워... 5쪽 쓸게 없는데 ㅡㅡ #P02#: 몬냐몬냐너가더잘써 ㅎㅎ #P01#: 5쪽 대충 의식의 흐름대로 쭉 써야지..이제 1쪽씀 ;; 5쪽 에는 네줄만 적어야지 #P02#: 안대... 뭔가분량중요할거같아 거의꽉채워서쓰셈 #P01#: 못써 쓸말업써 #P02#: 이거중간대체여?? #P01#: ㄴㄴ 그냥 과제임 그래서 더 짜증남'''

formatted_prompt = prompt_template.format(dialogue=dialogue)

# 토크나이징
inputs = tokenizer(
    formatted_prompt,
    return_tensors="pt"
).to("cuda")

outputs = model.generate(
    **inputs,
    max_new_tokens = 128,
    eos_token_id=tokenizer.eos_token_id, # EOS 토큰을 사용하여 명시적으로 출력의 끝을 지정.
    use_cache = True
)
decoded_outputs = tokenizer.batch_decode(outputs, skip_special_tokens=True)
result = decoded_outputs[0]

print(result)
result = result.split('### 요약 >>>')[-1].strip()
print(result)
```

This llama model was trained 2x faster with [Unsloth](https://github.com/unslothai/unsloth) and Huggingface's TRL library.

[<img src="https://raw.githubusercontent.com/unslothai/unsloth/main/images/unsloth%20made%20with%20love.png" width="200"/>](https://github.com/unslothai/unsloth)
I highly recommend checking the Unsloth notebook.