README.md

---
license: apache-2.0
tags:
- moe
- frankenmoe
- merge
- mergekit
- lazymergekit
- mlabonne/NeuralBeagle14-7B
- timpal0l/Mistral-7B-v0.1-flashback-v2
- Nexusflow/Starling-LM-7B-beta
- AI-Sweden-Models/tyr
base_model:
- mlabonne/NeuralBeagle14-7B
- timpal0l/Mistral-7B-v0.1-flashback-v2
- Nexusflow/Starling-LM-7B-beta
- AI-Sweden-Models/tyr
---

# MoEnsterBeagle

MoEnsterBeagle is a Mixture of Experts (MoE) made with the following models using [LazyMergekit](https://colab.research.google.com/drive/1obulZ1ROXHjYLn6PPZJwRR6GzgQogxxb?usp=sharing):
* [mlabonne/NeuralBeagle14-7B](https://huggingface.co/mlabonne/NeuralBeagle14-7B)
* [timpal0l/Mistral-7B-v0.1-flashback-v2](https://huggingface.co/timpal0l/Mistral-7B-v0.1-flashback-v2)
* [Nexusflow/Starling-LM-7B-beta](https://huggingface.co/Nexusflow/Starling-LM-7B-beta)
* [AI-Sweden-Models/tyr](https://huggingface.co/AI-Sweden-Models/tyr)

## 🧩 Configuration

```yaml
base_model: mlabonne/NeuralBeagle14-7B
gate_mode: cheap_embed
experts:
  - source_model: mlabonne/NeuralBeagle14-7B
    positive_prompts:
    - "chat"
    - "assistant"
    - "explain"
    - "tell me"
    - "english"
  - source_model: timpal0l/Mistral-7B-v0.1-flashback-v2
    positive_prompts:
    - "förklara"
    - "sammanfatta"
    - "svenska"
  - source_model: Nexusflow/Starling-LM-7B-beta
    positive_prompts: 
    - "code"
    - "programming"
    - "algorithm"
  - source_model: AI-Sweden-Models/tyr
    positive_prompts: 
    - "varför"
    - "förenkla"
    - "lagen"
```

## 💻 Usage

```python
!pip install -qU transformers bitsandbytes accelerate

from transformers import AutoTokenizer
import transformers
import torch

model = "FredrikBL/MoEnsterBeagle"

tokenizer = AutoTokenizer.from_pretrained(model)
pipeline = transformers.pipeline(
    "text-generation",
    model=model,
    model_kwargs={"torch_dtype": torch.float16, "load_in_4bit": True},
)

messages = [{"role": "user", "content": "Explain what a Mixture of Experts is in less than 100 words."}]
prompt = pipeline.tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
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"])
```