File size: 2,086 Bytes
6285e4f
 
 
362f552
 
 
6285e4f
 
 
e11b5cc
6285e4f
2106d66
6285e4f
 
 
 
 
 
 
 
 
 
 
 
2106d66
6285e4f
2d67bf4
e11b5cc
 
 
6285e4f
 
 
2106d66
6285e4f
 
 
2106d66
6285e4f
 
 
 
 
2106d66
6285e4f
2106d66
6285e4f
 
2106d66
6285e4f
 
 
 
 
 
2106d66
6285e4f
 
 
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
---
language:
- en
license: apache-2.0
tags:
- moe
inference: false
---
# Model Card for TinyMixtral-x8-Clonebase-7b
This model is based on [TinyLlama-1.1B](https://huggingface.co./TinyLlama/TinyLlama-1.1B-intermediate-step-1431k-3T), converted to a mistral model, and then placed the clone in mixtral.  
**This model was created experimentally for training a small mixtral.**  
**Without Train, the performance of this model is the same as TinyLlama.**

# How it was made
First, since tinyllama is an llama model, I converted it to a mistral model.  
  
After that, I cloned the FFN part and made it experts.
Since they are all the same tensor, the performance does not change.
All gates have the same value.

# How To Convert
use colab cpu-high-memory.  
This model was created with experts=8, but since it is a clone, you can create as many experts as you like.  

[tinyllama_to_mixtral_clonebase.ipynb](https://huggingface.co./mmnga/TinyMixtral-x8-Clonebase-7b/blob/main/notebook/tinyllama_to_mixtral_clonebase.ipynb)

# revision
[main TinyLlama-1.1B-intermediate-step-1431k-3T](https://huggingface.co./mmnga/TinyMixtral-x8-Clonebase-7b)  
[old  TinyLlama-1.1B-intermediate-step-1195k-token-2.5T](https://huggingface.co./mmnga/TinyMixtral-x8-Clonebase-7b/tree/TinyLlama-1.1B-intermediate-step-1195k-token-2.5T)

# Usage
~~~python
pip install transformers --upgrade
pip install flash_attn bitsandbytes accelerate
~~~

~~~python
from transformers import AutoTokenizer, AutoModelForCausalLM
import torch

model_name_or_path = "mmnga/TinyMixtral-x8-Clonebase-7b"

tokenizer = AutoTokenizer.from_pretrained(model_name_or_path)
model = AutoModelForCausalLM.from_pretrained(model_name_or_path, device_map="auto", load_in_8bit=True)

prompt = "Introducing the recipe for today's dinner."

with torch.no_grad():
    token_ids = tokenizer.encode(prompt, return_tensors="pt")
    output_ids = model.generate(
        token_ids.to(model.device),
        do_sample=True,
        max_new_tokens=128,
        repetition_penalty=1.5
    )
    output = tokenizer.decode(output_ids[0])
print(output)

~~~