File size: 2,054 Bytes
3a709f3
 
021d40a
 
 
 
 
 
 
 
 
 
 
 
 
3a709f3
021d40a
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
---
license: apache-2.0
tags:
- moe
- frankenmoe
- merge
- mergekit
- lazymergekit
- NousResearch/Nous-Hermes-2-Mistral-7B-DPO
- NeverSleep/Noromaid-7B-0.4-DPO
- mistralai/Mistral-7B-Instruct-v0.2
base_model:
- NousResearch/Nous-Hermes-2-Mistral-7B-DPO
- NeverSleep/Noromaid-7B-0.4-DPO
- mistralai/Mistral-7B-Instruct-v0.2
---

# Noro-Hermes-3x7B

Noro-Hermes-3x7B is a Mixture of Experts (MoE) made with the following models using [LazyMergekit](https://colab.research.google.com/drive/1obulZ1ROXHjYLn6PPZJwRR6GzgQogxxb?usp=sharing):
* [NousResearch/Nous-Hermes-2-Mistral-7B-DPO](https://huggingface.co/NousResearch/Nous-Hermes-2-Mistral-7B-DPO)
* [NeverSleep/Noromaid-7B-0.4-DPO](https://huggingface.co/NeverSleep/Noromaid-7B-0.4-DPO)
* [mistralai/Mistral-7B-Instruct-v0.2](https://huggingface.co/mistralai/Mistral-7B-Instruct-v0.2)

## 🧩 Configuration

```yaml
base_model: mistralai/Mistral-7B-Instruct-v0.2
dtype: bfloat16
gate_mode: hidden
experts:
  - source_model: NousResearch/Nous-Hermes-2-Mistral-7B-DPO
    positive_prompts: ["You are an intelligent assistant."]
  - source_model: NeverSleep/Noromaid-7B-0.4-DPO
    positive_prompts: ["You are a creative roleplaying assistant."]
  - source_model: mistralai/Mistral-7B-Instruct-v0.2
    positive_prompts: ["You are a general purpose assistant."]
```

## 💻 Usage

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

from transformers import AutoTokenizer
import transformers
import torch

model = "ThomasComics/Noro-Hermes-3x7B"

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"])
```