File size: 1,385 Bytes
e1124b4
9e67111
 
 
 
 
 
e1124b4
 
9e67111
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
b932ecc
9e67111
b932ecc
9e67111
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
---
language:
- ko
datasets:
- kyujinpy/KOR-Orca-Platypus-kiwi
library_name: transformers
pipeline_tag: text-generation
license: cc-by-nc-sa-4.0
---
**(주)미디어그룹사람과숲과 (주)마커의 LLM 연구 컨소시엄에서 개발된 모델입니다**  
**The license is `cc-by-nc-sa-4.0`.**  

# **KOR-Orca-Platypus-kiwi🥝**   

## Model Details

**Model Developers** Kyujin Han (kyujinpy)

**Model Architecture**  
ko-platypus-kiwi-13B is an auto-regressive language model based on the LLaMA2 transformer architecture.

**Base Model**  [hyunseoki/ko-en-llama2-13b](https://huggingface.co/hyunseoki/ko-en-llama2-13b)   

**Training Dataset**  
I used [kyujinpy/KOR-Orca-Platypus-kiwi](https://huggingface.co/datasets/kyujinpy/KOR-Orca-Platypus-kiwi).  


# Model comparisons
| Model | Average | Ko-ARC | Ko-HellaSwag | Ko-MMLU | Ko-TruthfulQA | Ko-CommonGen V2 |
| --- | --- | --- | --- | --- | --- | --- |
| **ko-platypus-kiwi-13B🥝** | 48.97 | 42.41 | 54.29 | 41.98 | 40.05 | **66.12** |  

# Implementation Code
```python
### KO-Platypus
from transformers import AutoModelForCausalLM, AutoTokenizer
import torch

repo = "kyujinpy/ko-platypus-kiwi-13B"
OpenOrca = AutoModelForCausalLM.from_pretrained(
        repo,
        return_dict=True,
        torch_dtype=torch.float16,
        device_map='auto'
)
OpenOrca_tokenizer = AutoTokenizer.from_pretrained(repo)
```

---