File size: 2,116 Bytes
5ff3702
 
 
 
 
 
 
 
 
 
 
935bc2e
5ff3702
 
 
 
 
 
935bc2e
5ff3702
 
 
 
 
 
 
 
 
 
 
 
 
935bc2e
5ff3702
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
63
64
---
license: apache-2.0
tags:
- moe
- frankenmoe
- merge
- mergekit
- lazymergekit
- Hemanth-thunder/Tamil-Mistral-7B-v0.1
base_model:
- Hemanth-thunder/Tamil-Mistral-7B-v0.1
- Hemanth-thunder/Tamil-Mistral-7B-v0.1
---

# SG-Tamil-MoE-7B

SG-Tamil-MoE-7B is a Mixture of Experts (MoE) made with the following models using [LazyMergekit](https://colab.research.google.com/drive/1obulZ1ROXHjYLn6PPZJwRR6GzgQogxxb?usp=sharing):
* [Hemanth-thunder/Tamil-Mistral-7B-v0.1](https://huggingface.co/Hemanth-thunder/Tamil-Mistral-7B-v0.1)
* [Hemanth-thunder/Tamil-Mistral-7B-v0.1](https://huggingface.co/Hemanth-thunder/Tamil-Mistral-7B-v0.1)

## 🧩 Configuration

```yaml
base_model: Hemanth-thunder/Tamil-Mistral-7B-v0.1
experts:
  - source_model: Hemanth-thunder/Tamil-Mistral-7B-v0.1
    positive_prompts:
    - "பேச்சு"  # "chat"
    - "உதவி"  # "assistant"
    - "எனக்கு சொல்"  # "tell me"
    - "விளக்கம்"  # "explain"
    - "நான் விரும்புகிறேன்"  # "I want"
  - source_model: Hemanth-thunder/Tamil-Mistral-7B-v0.1
    positive_prompts:
    - "அறிவுரை"  # "advice"
    - "நிர்வாகம்"  # "management"
    - "உத்தரவாதம்"  # "instructions"
    - "பயிற்சி"  # "training"
    - "செயல்முறை"  # "procedure"
```

## 💻 Usage

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

from transformers import AutoTokenizer
import transformers
import torch

model = "praveengovi/SG-Tamil-MoE-7B"

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"])
```