README.md

---
base_model:
- ssoyeun/Llama-3-BCCard-Kor-8B-dp
- ssoyeun/Llama-3-BCCard-Kor-8B-dp
- ssoyeun/Llama-3-BCCard-Kor-8B-dp
- ssoyeun/Llama-3-BCCard-Kor-8B-dp
- ssoyeun/Llama-3-BCCard-Kor-8B-dp
tags:
- merge
- mergekit
- lazymergekit
- ssoyeun/Llama-3-BCCard-Kor-8B-dp
---

# Llama-3-Kor-Bllossom-12B-psy-frankenstein

Llama-3-Kor-Bllossom-12B-psy-frankenstein is a merge of the following models using [LazyMergekit](https://colab.research.google.com/drive/1obulZ1ROXHjYLn6PPZJwRR6GzgQogxxb?usp=sharing):
* [ssoyeun/Llama-3-BCCard-Kor-8B-dp](https://huggingface.co/ssoyeun/Llama-3-BCCard-Kor-8B-dp)
* [ssoyeun/Llama-3-BCCard-Kor-8B-dp](https://huggingface.co/ssoyeun/Llama-3-BCCard-Kor-8B-dp)
* [ssoyeun/Llama-3-BCCard-Kor-8B-dp](https://huggingface.co/ssoyeun/Llama-3-BCCard-Kor-8B-dp)
* [ssoyeun/Llama-3-BCCard-Kor-8B-dp](https://huggingface.co/ssoyeun/Llama-3-BCCard-Kor-8B-dp)
* [ssoyeun/Llama-3-BCCard-Kor-8B-dp](https://huggingface.co/ssoyeun/Llama-3-BCCard-Kor-8B-dp)

## 🧩 Configuration

```yaml
slices:
  - sources:
    - model: ssoyeun/Llama-3-BCCard-Kor-8B-dp
      layer_range: [0,9]
  - sources:
    - model: ssoyeun/Llama-3-BCCard-Kor-8B-dp
      layer_range: [5,14]
  - sources:
    - model: ssoyeun/Llama-3-BCCard-Kor-8B-dp
      layer_range: [10,19]
  - sources:
    - model: ssoyeun/Llama-3-BCCard-Kor-8B-dp
      layer_range: [15,24]
  - sources:
    - model: ssoyeun/Llama-3-BCCard-Kor-8B-dp
      layer_range: [18,32]
merge_method: passthrough
dtype: bfloat16
```

## 💻 Usage

```python
!pip install -qU transformers accelerate

from transformers import AutoTokenizer
import transformers
import torch

model = "ssoyeun/Llama-3-Kor-Bllossom-12B-psy-frankenstein"
messages = [{"role": "user", "content": "What is a large language model?"}]

tokenizer = AutoTokenizer.from_pretrained(model)
prompt = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
pipeline = transformers.pipeline(
    "text-generation",
    model=model,
    torch_dtype=torch.float16,
    device_map="auto",
)

outputs = pipeline(prompt, max_new_tokens=256, do_sample=True, temperature=0.7, top_k=50, top_p=0.95)
print(outputs[0]["generated_text"])
```