File size: 1,278 Bytes
2f04424
 
164f3d6
 
 
 
2f04424
164f3d6
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
---
license: cc-by-nc-4.0
language:
- ko
library_name: transformers
pipeline_tag: text-generation
---
**The license is `cc-by-nc-4.0`.**  
  
# **GAI-LLM/OPEN-SOLAR-KO-10.7B-mixed-v15**  

## Model Details

**Model Developers** Donghoon Oh, Hanmin Myung, Eunyoung Kim (SK C&C G.AI Eng)

**Input** Models input text only.

**Output** Models generate text only.

**Model Architecture**  
GAI-LLM/OPEN-SOLAR-KO-10.7B-mixed-v15 is an auto-regressive language model based on the LLaMA2 transformer architecture.

**Base Model**  [beomi/OPEN-SOLAR-KO-10.7B](https://huggingface.co/beomi/OPEN-SOLAR-KO-10.7B)

**Training Dataset**  

- We combined Open Korean Dateset using mixed-strategy
- We use A100 GPU 80GB * 8, when training.

# **Model Benchmark**

## KO-LLM leaderboard
- Follow up as [Open KO-LLM LeaderBoard](https://huggingface.co/spaces/upstage/open-ko-llm-leaderboard).  

  
# Implementation Code
```python
### GAI-LLM/OPEN-SOLAR-KO-10.7B-mixed-v15
from transformers import AutoModelForCausalLM, AutoTokenizer
import torch

repo = "GAI-LLM/OPEN-SOLAR-KO-10.7B-mixed-v15"
model = AutoModelForCausalLM.from_pretrained(
        repo,
        return_dict=True,
        torch_dtype=torch.float16,
        device_map='auto'
)
tokenizer = AutoTokenizer.from_pretrained(repo)
```