---
license: apache-2.0
tags:
- moe
- merge
- mergekit
- NotAiLOL/Med-Yi-1.5-9B
- NotAiLOL/Yi-1.5-dolphin-9B
- NotAiLOL/Yi-1.5-9B-coder
- NotAiLOL/Math-Yi-1.5-9B
base_model:
- NotAiLOL/Med-Yi-1.5-9B
- NotAiLOL/Yi-1.5-dolphin-9B
- NotAiLOL/Yi-1.5-9B-coder
- NotAiLOL/Math-Yi-1.5-9B
---

# Eclipse-Yi-1.5-4x9b

Eclipse-Yi-1.5-4x9b is a Mixture of Experts (MoE) made with the following models:
* [NotAiLOL/Med-Yi-1.5-9B](https://huggingface.co/NotAiLOL/Med-Yi-1.5-9B)
* [NotAiLOL/Yi-1.5-dolphin-9B](https://huggingface.co/NotAiLOL/Yi-1.5-dolphin-9B)
* [NotAiLOL/Yi-1.5-9B-coder](https://huggingface.co/NotAiLOL/Yi-1.5-9B-coder)
* [NotAiLOL/Math-Yi-1.5-9B](https://huggingface.co/NotAiLOL/Math-Yi-1.5-9B)

## 🧩 Configuration

```yaml
base_model: 01-ai/Yi-1.5-9B
gate_mode: hidden
dtype: bfloat16
experts:
    - source_model: NotAiLOL/Med-Yi-1.5-9B
      positive_prompts:
      - "explain"
      - "write"
    - source_model: NotAiLOL/Yi-1.5-dolphin-9B
      positive_prompts:
      - "chat"
      - "assistant"
      - "tell me"
      - "explain"
      - "I want"
    - source_model: NotAiLOL/Yi-1.5-9B-coder
      positive_prompts:
      - "chat"
      - "assistant"
      - "tell me"
      - "explain"
      - "I want"
    - source_model: NotAiLOL/Math-Yi-1.5-9B
      positive_prompts:
      - "reason"
      - "math"
      - "mathematics"
      - "solve"
      - "count"
```

## 💻 Usage

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

from transformers import AutoTokenizer
import transformers
import torch

model = "NotAiLOL/Eclipse-Yi-1.5-4x9b"

tokenizer = AutoTokenizer.from_pretrained(model)
pipeline = transformers.pipeline(
    "text-generation",
    model=model,
    model_kwargs={"torch_dtype": torch.bfloat16, "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"])
```