Wiedervereinigung-7b-dpo/README.md

---
tags:
- merge
- mergekit
- lazymergekit
- DiscoResearch/DiscoLM_German_7b_v1
- DRXD1000/Phoenix
- VAGOsolutions/SauerkrautLM-7b-v1-mistral
- malteos/hermeo-7b
base_model:
- DiscoResearch/DiscoLM_German_7b_v1
- DRXD1000/Phoenix
- VAGOsolutions/SauerkrautLM-7b-v1-mistral
- malteos/hermeo-7b
license: apache-2.0
language:
- de
- en
---

# Wiedervereinigung-7b-dpo
![image/png](https://huggingface.co/mayflowergmbh/Wiedervereinigung-7b/resolve/main/Wiedervereinigung-7b.png)

This is a dpo aligned merge of our favourite german models, scoring 7.11 on the mt-bench-de average.
Since the original models based on mistral - three of them on the brilliant german LeoLM/leo-mistral-hessianai-7b - they are reunited in this merged model. 
Therefore the name, no nationalist ideas involved :-). 

To improve result quality they are dpo-trained with a german translation of slimorca dpo using hermeo-7B for reject results. 

If you are gpu-poor like me you can now use [LLaMA-Factory](https://github.com/hiyouga/LLaMA-Factory) to train with german datasets.

Kudos to the authors of the original models at [DiscoResearch](https://huggingface.co/DiscoResearch) and [VAGOsolutions](https://huggingface.co/VAGOsolutions), [Malte Ostendorff](https://huggingface.co/malteos) 
and [Matthias Uhlig](https://huggingface.co/DRXD1000). We are your fan club.

This model was brought to you and the nvidia bill was paid by [Mayflower GmbH](https://mayflower.de/).

## Benchmark results: mt-bench-de

Is the merged model alone already good? Well, of course. But it is even better with the help of some dpo tuning.

```json
{
    "first_turn": 7.3,
    "second_turn": 6.925,
    "categories": {
        "writing": 8.425,
        "roleplay": 8.6,
        "reasoning": 5.4,
        "math": 4.35,
        "coding": 4.3,
        "extraction": 7.975,
        "stem": 8.5,
        "humanities": 9.35
    },
    "average": 7.1125
}
```

## Other Versions

A big thank you to [LoneStriker](https://huggingface.co/LoneStriker) for the quantized models.

| Name | Quant method | Bits | 
| ---- | ---- | ---- | 
[Wiedervereinigung-7b-dpo](https://huggingface.co/mayflowergmbh/Wiedervereinigung-7b-dpo)| Unquantized | 16 |
[Wiedervereinigung-7b-dpo-GPTQ](https://huggingface.co/LoneStriker/Wiedervereinigung-7b-dpo-GPTQ)| GPTQ | 4 |
[Wiedervereinigung-7b-dpo-AWQ](https://huggingface.co/LoneStriker/Wiedervereinigung-7b-dpo-AWQ)| AWQ | 4 |
[Wiedervereinigung-7b-dpo-GGUF](https://huggingface.co/LoneStriker/Wiedervereinigung-7b-dpo-GGUF)| GGUF | 3-8 |
[Wiedervereinigung-7b-dpo-8.0bpw-h8-exl2](https://huggingface.co/LoneStriker/Wiedervereinigung-7b-dpo-8.0bpw-h8-exl2)| EXL2 | 8 |
[Wiedervereinigung-7b-dpo-6.0bpw-h6-exl2](https://huggingface.co/LoneStriker/Wiedervereinigung-7b-dpo-6.0bpw-h6-exl2)| EXL2 | 6 |
[Wiedervereinigung-7b-dpo-5.0bpw-h6-exl2](https://huggingface.co/LoneStriker/Wiedervereinigung-7b-dpo-5.0bpw-h6-exl2)| EXL2 | 5 |
[Wiedervereinigung-7b-dpo-4.0bpw-h6-exl2](https://huggingface.co/LoneStriker/Wiedervereinigung-7b-dpo-4.0bpw-h6-exl2)| EXL2 | 4 |
[Wiedervereinigung-7b-dpo-3.0bpw-h6-exl2](https://huggingface.co/LoneStriker/Wiedervereinigung-7b-dpo-3.0bpw-h6-exl2)| EXL2 | 3 |

Wiedervereinigung-7b is a  [LazyMergekit](https://colab.research.google.com/drive/1obulZ1ROXHjYLn6PPZJwRR6GzgQogxxb?usp=sharing) merge of:
* [DiscoResearch/DiscoLM_German_7b_v1](https://huggingface.co/DiscoResearch/DiscoLM_German_7b_v1)
* [DRXD1000/Phoenix](https://huggingface.co/DRXD1000/Phoenix)
* [VAGOsolutions/SauerkrautLM-7b-v1-mistral](https://huggingface.co/VAGOsolutions/SauerkrautLM-7b-v1-mistral)
* [malteos/hermeo-7b](https://huggingface.co/malteos/hermeo-7b)


## 🧩 Configuration

```yaml
models:
  - model: LeoLM/leo-mistral-hessianai-7b
    # No parameters necessary for base model
  - model: DiscoResearch/DiscoLM_German_7b_v1
    parameters:
      density: 0.6
      weight: 0.25
  - model: DRXD1000/Phoenix
    parameters:
      density: 0.6
      weight: 0.25
  - model: VAGOsolutions/SauerkrautLM-7b-v1-mistral
    parameters:
      density: 0.6
      weight: 0.25
  - model: malteos/hermeo-7b
    parameters:
      density: 0.6
      weight: 0.25
merge_method: dare_ties
base_model: LeoLM/leo-mistral-hessianai-7b
parameters:
  int8_mask: true
dtype: bfloat16
```


## 💻 Usage

```python
!pip install -qU transformers accelerate

from transformers import AutoTokenizer
import transformers
import torch

model = "mayflowergmbh/Wiedervereinigung-7b-dpo"
messages = [{"role": "user", "content": "Was ist ein deutsches 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"])
```