kyujinpy commited on
Commit
0358880
1 Parent(s): a7466c0

Upload 2 files

Browse files
Files changed (3) hide show
  1. .gitattributes +1 -0
  2. Korean-OpenOrca.png +3 -0
  3. README.md +64 -0
.gitattributes CHANGED
@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
 
 
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
36
+ Korean-OpenOrca.png filter=lfs diff=lfs merge=lfs -text
Korean-OpenOrca.png ADDED

Git LFS Details

  • SHA256: f25dd74abb56027e8e0bc12085ee4b8ce6a374d311993d997f84164709dc8003
  • Pointer size: 132 Bytes
  • Size of remote file: 3.91 MB
README.md CHANGED
@@ -1,3 +1,67 @@
1
  ---
 
 
 
 
 
 
2
  license: cc-by-nc-sa-4.0
3
  ---
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
  ---
2
+ language:
3
+ - ko
4
+ datasets:
5
+ - kyujinpy/KOR-OpenOrca-Platypus-v3
6
+ library_name: transformers
7
+ pipeline_tag: text-generation
8
  license: cc-by-nc-sa-4.0
9
  ---
10
+ **(주)미디어그룹사람과숲과 (주)마커의 LLM 연구 컨소시엄에서 개발된 모델입니다**
11
+ **The license is `cc-by-nc-sa-4.0`.**
12
+
13
+ # **🐳Korean-OpenOrca-13B🐳**
14
+ ![img](./Korean-OpenOrca.png)
15
+
16
+ ## Model Details
17
+
18
+ **Model Developers** Kyujin Han (kyujinpy)
19
+
20
+ **Input** Models input text only.
21
+
22
+ **Output** Models generate text only.
23
+
24
+ **Model Architecture**
25
+ Korean-OpenOrca-13B is an auto-regressive language model based on the LLaMA2 transformer architecture.
26
+
27
+ **Repo Link**
28
+ Github Korean-OpenOrca: [🐳Korean-OpenOrca🐳](https://github.com/Marker-Inc-Korea/Korean-OpenOrca)
29
+
30
+ **Base Model** [hyunseoki/ko-en-llama2-13b](https://huggingface.co/hyunseoki/ko-en-llama2-13b)
31
+
32
+ **Training Dataset**
33
+ I use [kyujinpy/KOR-OpenOrca-Platypus-v3(private! wait!)](https://huggingface.co/datasets/kyujinpy/KOR-OpenOrca-Platypus-v3).
34
+
35
+ I use A100 GPU 40GB and COLAB, when trianing.
36
+
37
+
38
+ # **Model Benchmark**
39
+
40
+ ## KO-LLM leaderboard
41
+ - Follow up as [Open KO-LLM LeaderBoard](https://huggingface.co/spaces/upstage/open-ko-llm-leaderboard).
42
+
43
+ | Model | Average |Ko-ARC | Ko-HellaSwag | Ko-MMLU | Ko-TruthfulQA | Ko-CommonGen V2 |
44
+ | --- | --- | --- | --- | --- | --- | --- |
45
+ | [Korean-OpenOrca-13B] | 46.59 | 42.06 | 53.95 | 42.28 | 43.55 | 51.12 |
46
+ | Korean-OpenOrca-13B-v2 | 49.48 | 44.03 | 54.43 | 42.23 | 41.64 | 65.05 |
47
+
48
+ > Compare with Top 4 SOTA models. (update: 10/09)
49
+
50
+
51
+ # Implementation Code
52
+ ```python
53
+ ### KO-Platypus
54
+ from transformers import AutoModelForCausalLM, AutoTokenizer
55
+ import torch
56
+
57
+ repo = "kyujinpy/Korean-OpenOrca-13B-v2"
58
+ OpenOrca = AutoModelForCausalLM.from_pretrained(
59
+ repo,
60
+ return_dict=True,
61
+ torch_dtype=torch.float16,
62
+ device_map='auto'
63
+ )
64
+ OpenOrca_tokenizer = AutoTokenizer.from_pretrained(repo)
65
+ ```
66
+
67
+ ---