Hugging Face's logo

kyujinpy
/
KO-Platypus2-7B-ex

Text Generation
Transformers
PyTorch
English
Korean
llama
text-generation-inference
Inference Endpoints
Model card Files Community
1
kyujinpy commited on
Commit
b7a6f9d
·
1 Parent(s): bb1898f

Upload README.md

Browse files
Files changed (1) hide show
  1. README.md +17 -9
README.md CHANGED
@@ -73,8 +73,10 @@ I use A100 GPU 40GB and COLAB, when trianing.
73
  | [Polyglot-ko-12.8b](https://huggingface.co/EleutherAI/polyglot-ko-12.8b) | 0.7937 | 0.8108 | 0.8037 | 0.8369 |
74
  | [Llama-2-Ko-7b 20B](https://huggingface.co/beomi/llama-2-ko-7b) | 0.7388 | 0.7626 | 0.7808 | 0.7979 |
75
  | [Llama-2-Ko-7b 40B](https://huggingface.co/beomi/llama-2-ko-7b) | 0.7436 | 0.7927 | 0.8037 | 0.8259 |
76
- | [KO-platypus2-13B](https://huggingface.co/kyujinpy/KO-Platypus2-13B) | 0.5820 | 0.6269 | 0.6267 | 0.6527 |
77
- | **KO-platypus2-7B-EX(ours)** | 0.7509 | 0.7899 | 0.8029 | 0.8290 |
 
 
78
 
79
  > Natural Language Inference (NLI; 자연어 추론 평가)
80
  ### HellaSwag (F1)
@@ -87,8 +89,10 @@ I use A100 GPU 40GB and COLAB, when trianing.
87
  | [Polyglot-ko-12.8b](https://huggingface.co/EleutherAI/polyglot-ko-12.8b) | 0.5954 | 0.6306 | 0.6098 | 0.6118 |
88
  | [Llama-2-Ko-7b 20B](https://huggingface.co/beomi/llama-2-ko-7b) | 0.4518 | 0.4668 | 0.4726 | 0.4828 |
89
  | [Llama-2-Ko-7b 40B](https://huggingface.co/beomi/llama-2-ko-7b) | 0.4562 | 0.4657 | 0.4698 | 0.4774 |
90
- | [KO-platypus2-13B](https://huggingface.co/kyujinpy/KO-Platypus2-13B) | 0.3912 | 0.4129 | 0.4144 | 0.4330 |
91
- | **KO-platypus2-7B-EX(ours)** | 0.4571 | 0.4461 | 0.4371 | 0.4525 |
 
 
92
 
93
  > Question Answering (QA)
94
  ### BoolQ (F1)
@@ -101,8 +105,10 @@ I use A100 GPU 40GB and COLAB, when trianing.
101
  | [Polyglot-ko-12.8b](https://huggingface.co/EleutherAI/polyglot-ko-12.8b) | 0.4818 | 0.6041 | 0.6289 | 0.6448 |
102
  | [Llama-2-Ko-7b 20B](https://huggingface.co/beomi/llama-2-ko-7b) | 0.3607 | 0.6797 | 0.6801 | 0.6622 |
103
  | [Llama-2-Ko-7b 40B](https://huggingface.co/beomi/llama-2-ko-7b) | 0.5786 | 0.6977 | 0.7084 | 0.7144 |
104
- | [KO-platypus2-13B](https://huggingface.co/kyujinpy/KO-Platypus2-13B) | 0.3539 | 0.7168 | 0.7328 | 0.7172 |
105
- | **KO-platypus2-7B-EX(ours)** | 0.6028 | 0.6979 | 0.7016 | 0.6988 |
 
 
106
 
107
  > Classification
108
  ### SentiNeg (F1)
@@ -115,9 +121,11 @@ I use A100 GPU 40GB and COLAB, when trianing.
115
  | [Polyglot-ko-12.8b](https://huggingface.co/EleutherAI/polyglot-ko-12.8b) | 0.9117 | 0.9015 | 0.9345 | 0.9723 |
116
  | [Llama-2-Ko-7b 20B](https://huggingface.co/beomi/llama-2-ko-7b) | 0.4855 | 0.8295 | 0.8711 | 0.8513 |
117
  | [Llama-2-Ko-7b 40B](https://huggingface.co/beomi/llama-2-ko-7b) | 0.4594 | 0.7611 | 0.7276 | 0.9370 |
118
- | [KO-platypus2-13B](https://huggingface.co/kyujinpy/KO-Platypus2-13B) | 0.5216 | 0.8236 | 0.8487 | 0.8789 |
119
- | **KO-platypus2-7B-EX(ours)** | 0.5821 | 0.7653 | 0.7991 | 0.8643 |
120
-
 
 
121
  # Implementation Code
122
  ```python
123
  ### KO-Platypus
 
73
  | [Polyglot-ko-12.8b](https://huggingface.co/EleutherAI/polyglot-ko-12.8b) | 0.7937 | 0.8108 | 0.8037 | 0.8369 |
74
  | [Llama-2-Ko-7b 20B](https://huggingface.co/beomi/llama-2-ko-7b) | 0.7388 | 0.7626 | 0.7808 | 0.7979 |
75
  | [Llama-2-Ko-7b 40B](https://huggingface.co/beomi/llama-2-ko-7b) | 0.7436 | 0.7927 | 0.8037 | 0.8259 |
76
+ | [*Platypus2-7B](https://huggingface.co/garage-bAInd/Platypus2-7B) | NaN | NaN | NaN | NaN |
77
+ | **KO-platypus2-7B-EX(ours)** | 0.7509 | 0.7899 | 0.8029 | 0.8290 |
78
+ *Platypus2-7B Original used https://huggingface.co/garage-bAInd/Platypus2-7B
79
+ *Platypus2-13B Original used https://huggingface.co/garage-bAInd/Platypus2-13B
80
 
81
  > Natural Language Inference (NLI; 자연어 추론 평가)
82
  ### HellaSwag (F1)
 
89
  | [Polyglot-ko-12.8b](https://huggingface.co/EleutherAI/polyglot-ko-12.8b) | 0.5954 | 0.6306 | 0.6098 | 0.6118 |
90
  | [Llama-2-Ko-7b 20B](https://huggingface.co/beomi/llama-2-ko-7b) | 0.4518 | 0.4668 | 0.4726 | 0.4828 |
91
  | [Llama-2-Ko-7b 40B](https://huggingface.co/beomi/llama-2-ko-7b) | 0.4562 | 0.4657 | 0.4698 | 0.4774 |
92
+ | [*Platypus2-7B](https://huggingface.co/garage-bAInd/Platypus2-7B) | NaN | NaN | NaN | NaN |
93
+ | **KO-platypus2-7B-EX(ours)** | 0.4571 | 0.4461 | 0.4371 | 0.4525 |
94
+ *Platypus2-7B Original used https://huggingface.co/garage-bAInd/Platypus2-7B
95
+ *Platypus2-13B Original used https://huggingface.co/garage-bAInd/Platypus2-13B
96
 
97
  > Question Answering (QA)
98
  ### BoolQ (F1)
 
105
  | [Polyglot-ko-12.8b](https://huggingface.co/EleutherAI/polyglot-ko-12.8b) | 0.4818 | 0.6041 | 0.6289 | 0.6448 |
106
  | [Llama-2-Ko-7b 20B](https://huggingface.co/beomi/llama-2-ko-7b) | 0.3607 | 0.6797 | 0.6801 | 0.6622 |
107
  | [Llama-2-Ko-7b 40B](https://huggingface.co/beomi/llama-2-ko-7b) | 0.5786 | 0.6977 | 0.7084 | 0.7144 |
108
+ | [*Platypus2-7B](https://huggingface.co/garage-bAInd/Platypus2-7B) | NaN | NaN | NaN | NaN |
109
+ | **KO-platypus2-7B-EX(ours)** | 0.6028 | 0.6979 | 0.7016 | 0.6988 |
110
+ *Platypus2-7B Original used https://huggingface.co/garage-bAInd/Platypus2-7B
111
+ *Platypus2-13B Original used https://huggingface.co/garage-bAInd/Platypus2-13B
112
 
113
  > Classification
114
  ### SentiNeg (F1)
 
121
  | [Polyglot-ko-12.8b](https://huggingface.co/EleutherAI/polyglot-ko-12.8b) | 0.9117 | 0.9015 | 0.9345 | 0.9723 |
122
  | [Llama-2-Ko-7b 20B](https://huggingface.co/beomi/llama-2-ko-7b) | 0.4855 | 0.8295 | 0.8711 | 0.8513 |
123
  | [Llama-2-Ko-7b 40B](https://huggingface.co/beomi/llama-2-ko-7b) | 0.4594 | 0.7611 | 0.7276 | 0.9370 |
124
+ | [*Platypus2-7B](https://huggingface.co/garage-bAInd/Platypus2-7B) | NaN | NaN | NaN | NaN |
125
+ | **KO-platypus2-7B-EX(ours)** | 0.5821 | 0.7653 | 0.7991 | 0.8643 |
126
+ *Platypus2-7B Original used https://huggingface.co/garage-bAInd/Platypus2-7B
127
+ *Platypus2-13B Original used https://huggingface.co/garage-bAInd/Platypus2-13B
128
+
129
  # Implementation Code
130
  ```python
131
  ### KO-Platypus