README.md

---
license: apache-2.0
tags:
- merge
- mergekit
- lazymergekit
- Cartinoe5930/Llama2_init_Mistral
- Cartinoe5930/Llama2_init_Mistral
base_model:
- Cartinoe5930/Llama2_init_Mistral
- Cartinoe5930/Llama2_init_Mistral
---

# SOLAR-DUS-implement

SOLAR-DUS-implement is a merge of the following model using [LazyMergekit](https://colab.research.google.com/drive/1obulZ1ROXHjYLn6PPZJwRR6GzgQogxxb?usp=sharing):
* [Cartinoe5930/Llama2_init_Mistral](https://huggingface.co/Cartinoe5930/Llama2_init_Mistral)

For more detailed information, please refer to GitHub Repository.

GitHub Repository: https://github.com/gauss5930/iDUS

## 🧩 Configuration

```yaml
slices:
  - sources:
    - model: Cartinoe5930/Llama2_init_Mistral
      layer_range: [0, 24]
  - sources:
    - model: Cartinoe5930/Llama2_init_Mistral
      layer_range: [8, 32]
merge_method: passthrough
dtype: float16
```

## 🏆 HuggingFace Open LLM Leaderboard

|Model|ARC|HellaSwag|MMLU|TruthfulQA|Winogrande|GSM8K|Average|
|---|---|---|---|---|---|---|---|
|SOLAR-10.7B-DUS-Implementation|59.56|81.18|63.68|40.72|76.48|26.99|58.1|

## 💻 Usage

```python
!pip install -qU transformers accelerate

from transformers import AutoTokenizer
import transformers
import torch

model = "Cartinoe5930/SOLAR-DUS-implement"
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"])
```