File size: 22,938 Bytes
9577768
101d48e
 
 
 
9577768
101d48e
 
 
 
 
 
 
 
 
 
 
 
9577768
101d48e
 
 
 
 
7d076c3
101d48e
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
ae6b2cc
101d48e
 
 
0960346
 
f5aff36
 
 
0960346
 
 
 
101d48e
 
04620d1
101d48e
 
 
 
 
 
 
 
 
 
 
 
 
04620d1
101d48e
04620d1
101d48e
 
 
 
 
 
 
 
 
 
 
04620d1
101d48e
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
0960346
 
f5aff36
 
 
0960346
 
 
 
101d48e
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
---
language:
- en
- ko
license: llama3
library_name: transformers
tags:
- translation
- enko
- ko
base_model:
- meta-llama/Meta-Llama-3-8B-Instruct
datasets:
- nayohan/aihub-en-ko-translation-1.2m
- nayohan/translate_corpus_313k
pipeline_tag: text-generation
metrics:
- sacrebleu
---
# **instructTrans**

![image/png](https://cdn-uploads.huggingface.co/production/uploads/6152b4b9ecf3ca6ab820e325/oRlzxHQy3Qvqf4zfh5Wcj.png)
# **Introduction**

**llama3-8b-instructTrans-en-ko** model is trained on **translation datasets(english->korean)** based on Llama-3-8B-it. To translate the English instruction dataset.
- [nayohan/aihub-en-ko-translation-1.2m](https://huggingface.co/datasets/nayohan/aihub-en-ko-translation-1.2m)
- [nayohan/translate_corpus_313k](https://huggingface.co/datasets/nayohan/translate_corpus_313k)



### **Loading the Model**
Use the following Python code to load the model:
```python
import torch
from transformers import AutoModelForCausalLM, AutoTokenizer

model_name = "nayohan/llama3-instrucTrans-enko-8b"
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForCausalLM.from_pretrained(
  model_name,
  device_map="auto",
  torch_dtype=torch.bfloat16
)
```

### **Generating Text**
This model supports translation from english to korean. To translate text, use the following Python code: 
```python
system_prompt="๋‹น์‹ ์€ ๋ฒˆ์—ญ๊ธฐ ์ž…๋‹ˆ๋‹ค. ์˜์–ด๋ฅผ ํ•œ๊ตญ์–ด๋กœ ๋ฒˆ์—ญํ•˜์„ธ์š”."
sentence = "The aerospace industry is a flower in the field of technology and science."
conversation = [{'role': 'system', 'content': system_prompt},
                {'role': 'user', 'content': sentence}]

inputs = tokenizer.apply_chat_template(
  conversation,
  tokenize=True,
  add_generation_prompt=True,
  return_tensors='pt'
).to("cuda")

outputs = model.generate(inputs, max_new_tokens=4096) # Finetuned with length 4096
print(tokenizer.decode(outputs[0][len(inputs[0]):]))
```
```
# Result
INPUT: <|begin_of_text|><|start_header_id|>system<|end_header_id|>\n\n๋‹น์‹ ์€ ๋ฒˆ์—ญ๊ธฐ ์ž…๋‹ˆ๋‹ค. ์˜์–ด๋ฅผ ํ•œ๊ตญ์–ด๋กœ ๋ฒˆ์—ญํ•˜์„ธ์š”.<|eot_id|><|start_header_id|>user<|end_header_id|>\n\nThe aerospace industry is a flower in the field of technology and science.<|eot_id|><|start_header_id|>assistant<|end_header_id|>\n\n
OUTPUT: ํ•ญ๊ณต์šฐ์ฃผ ์‚ฐ์—…์€ ๊ธฐ์ˆ ๊ณผ ๊ณผํ•™ ๋ถ„์•ผ์˜ ๊ฝƒ์ž…๋‹ˆ๋‹ค.<|eot_id|>

INPUT: <|begin_of_text|><|start_header_id|>system<|end_header_id|>\n\n๋‹น์‹ ์€ ๋ฒˆ์—ญ๊ธฐ ์ž…๋‹ˆ๋‹ค. ์˜์–ด๋ฅผ ํ•œ๊ตญ์–ด๋กœ ๋ฒˆ์—ญํ•˜์„ธ์š”.<|eot_id|><|start_header_id|>user<|end_header_id|>\n\n
Technical and basic sciences are very important in terms of research. It has a significant impact on the industrial development of a country. Government policies control the research budget.<|eot_id|><|start_header_id|>assistant<|end_header_id|>\n\n
OUTPUT: ๊ธฐ์ˆ  ๋ฐ ๊ธฐ์ดˆ ๊ณผํ•™์€ ์—ฐ๊ตฌ ์ธก๋ฉด์—์„œ ๋งค์šฐ ์ค‘์š”ํ•ฉ๋‹ˆ๋‹ค. ์ด๋Š” ํ•œ ๊ตญ๊ฐ€์˜ ์‚ฐ์—… ๋ฐœ์ „์— ํฐ ์˜ํ–ฅ์„ ๋ฏธ์นฉ๋‹ˆ๋‹ค. ์ •๋ถ€ ์ •์ฑ…์€ ์—ฐ๊ตฌ ์˜ˆ์‚ฐ์„ ํ†ต์ œํ•ฉ๋‹ˆ๋‹ค.<|eot_id|>
```
```
# EVAL_RESULT (2405_KO_NEWS) (max_new_tokens=512)
"en_ref":"This controversy arose around a new advertisement for the latest iPad Pro that Apple released on YouTube on the 7th. The ad shows musical instruments, statues, cameras, and paints being crushed in a press, followed by the appearance of the iPad Pro in their place. It appears to emphasize the new iPad Pro's artificial intelligence features, advanced display, performance, and thickness. Apple mentioned that the newly unveiled iPad Pro is equipped with the latest 'M4' chip and is the thinnest device in Apple's history. The ad faced immediate backlash upon release, as it graphically depicts objects symbolizing creators being crushed. Critics argue that the imagery could be interpreted as technology trampling on human creators. Some have also voiced concerns that it evokes a situation where creators are losing ground due to AI."
"ko_ref":"์ด๋ฒˆ ๋…ผ๋ž€์€ ์• ํ”Œ์ด ์ง€๋‚œ 7์ผ ์œ ํŠœ๋ธŒ์— ๊ณต๊ฐœํ•œ ์‹ ํ˜• ์•„์ดํŒจ๋“œ ํ”„๋กœ ๊ด‘๊ณ ๋ฅผ ๋‘˜๋Ÿฌ์‹ธ๊ณ  ๋ถˆ๊ฑฐ์กŒ๋‹ค. ํ•ด๋‹น ๊ด‘๊ณ  ์˜์ƒ์€ ์•…๊ธฐ์™€ ์กฐ๊ฐ์ƒ, ์นด๋ฉ”๋ผ, ๋ฌผ๊ฐ ๋“ฑ์„ ์••์ฐฉ๊ธฐ๋กœ ์ง“๋ˆ„๋ฅธ ๋’ค ๊ทธ ์ž๋ฆฌ์— ์•„์ดํŒจ๋“œ ํ”„๋กœ๋ฅผ ๋“ฑ์žฅ์‹œํ‚ค๋Š” ๋‚ด์šฉ์ด์—ˆ๋‹ค. ์‹ ํ˜• ์•„์ดํŒจ๋“œ ํ”„๋กœ์˜ ์ธ๊ณต์ง€๋Šฅ ๊ธฐ๋Šฅ๋“ค๊ณผ ์ง„ํ™”๋œ ๋””์Šคํ”Œ๋ ˆ์ด์™€ ์„ฑ๋Šฅ, ๋‘๊ป˜ ๋“ฑ์„ ๊ฐ•์กฐํ•˜๊ธฐ ์œ„ํ•œ ์ทจ์ง€๋กœ ํ’€์ด๋œ๋‹ค. ์• ํ”Œ์€ ์ด๋ฒˆ์— ๊ณต๊ฐœํ•œ ์•„์ดํŒจ๋“œ ํ”„๋กœ์— ์‹ ํ˜• โ€˜M4โ€™ ์นฉ์ด ํƒ‘์žฌ๋˜๋ฉฐ ๋‘๊ป˜๋Š” ์• ํ”Œ์˜ ์—ญ๋Œ€ ์ œํ’ˆ ์ค‘ ๊ฐ€์žฅ ์–‡๋‹ค๋Š” ์„ค๋ช…๋„ ๋ง๋ถ™์˜€๋‹ค. ๊ด‘๊ณ ๋Š” ๊ณต๊ฐœ ์งํ›„ ๊ฑฐ์„ผ ๋น„ํŒ์— ์ง๋ฉดํ–ˆ๋‹ค. ์ฐฝ์ž‘์ž๋ฅผ ์ƒ์ง•ํ•˜๋Š” ๋ฌผ๊ฑด์ด ์ง“๋ˆŒ๋ ค์ง€๋Š” ๊ณผ์ •์„ ์ง€๋‚˜์น˜๊ฒŒ ์ ๋‚˜๋ผํ•˜๊ฒŒ ๋ฌ˜์‚ฌํ•œ ์ ์ด ๋ฌธ์ œ๊ฐ€ ๋๋‹ค. ๊ธฐ์ˆ ์ด ์ธ๊ฐ„ ์ฐฝ์ž‘์ž๋ฅผ ์ง“๋ฐŸ๋Š” ๋ชจ์Šต์„ ๋ฌ˜์‚ฌํ•œ ๊ฒƒ์œผ๋กœ ํ•ด์„๋  ์—ฌ์ง€๊ฐ€ ์žˆ๋‹ค๋Š” ๋ฌธ์ œ์˜์‹์ด๋‹ค. ์ธ๊ณต์ง€๋Šฅ(AI)์œผ๋กœ ์ธํ•ด ์ฐฝ์ž‘์ž๊ฐ€ ์„ค ์ž๋ฆฌ๊ฐ€ ์ค„์–ด๋“œ๋Š” ์ƒํ™ฉ์„ ์—ฐ์ƒ์‹œํ‚จ๋‹ค๋Š” ๋ชฉ์†Œ๋ฆฌ๋„ ๋‚˜์™”๋‹ค."

"InstrucTrans":"์ด๋ฒˆ ๋…ผ๋ž€์€ ์• ํ”Œ์ด ์ง€๋‚œ 7์ผ ์œ ํŠœ๋ธŒ์— ๊ณต๊ฐœํ•œ ์ตœ์‹  ์•„์ดํŒจ๋“œ ํ”„๋กœ ๊ด‘๊ณ ๋ฅผ ์ค‘์‹ฌ์œผ๋กœ ๋ถˆ๊ฑฐ์กŒ๋‹ค. ์ด ๊ด‘๊ณ ๋Š” ์•…๊ธฐ, ์กฐ๊ฐ์ƒ, ์นด๋ฉ”๋ผ, ๋ฌผ๊ฐ ๋“ฑ์„ ๋ˆ„๋ฅด๊ธฐ ์‹œ์ž‘ํ•˜๋Š” ์žฅ๋ฉด๊ณผ ํ•จ๊ป˜ ๊ทธ ์ž๋ฆฌ์— ์•„์ดํŒจ๋“œ ํ”„๋กœ๊ฐ€ ๋“ฑ์žฅํ•˜๋Š” ์žฅ๋ฉด์„ ๋ณด์—ฌ์ค€๋‹ค. ์ด๋Š” ์ƒˆ๋กœ์šด ์•„์ดํŒจ๋“œ ํ”„๋กœ์˜ ์ธ๊ณต์ง€๋Šฅ ๊ธฐ๋Šฅ, ๊ณ ๊ธ‰ ๋””์Šคํ”Œ๋ ˆ์ด, ์„ฑ๋Šฅ, ๋‘๊ป˜๋ฅผ ๊ฐ•์กฐํ•˜๋Š” ๊ฒƒ์œผ๋กœ ๋ณด์ธ๋‹ค. ์• ํ”Œ์€ ์ด๋ฒˆ์— ๊ณต๊ฐœํ•œ ์•„์ดํŒจ๋“œ ํ”„๋กœ์— ์ตœ์‹  'M4' ์นฉ์ด ํƒ‘์žฌ๋์œผ๋ฉฐ, ์• ํ”Œ ์—ญ์‚ฌ์ƒ ๊ฐ€์žฅ ์–‡์€ ๊ธฐ๊ธฐ๋ผ๊ณ  ์–ธ๊ธ‰ํ–ˆ๋‹ค. ์ด ๊ด‘๊ณ ๋Š” ์ถœ์‹œํ•˜์ž๋งˆ์ž ํฌ๋ฆฌ์—์ดํ„ฐ๋ฅผ ์ƒ์ง•ํ•˜๋Š” ๋ฌผ๊ฑด์ด ํŒŒ์‡„๋˜๋Š” ์žฅ๋ฉด์ด ๊ทธ๋Œ€๋กœ ๊ทธ๋ ค์ ธ ๋…ผ๋ž€์ด ๋˜๊ณ  ์žˆ๋‹ค. ๋น„ํ‰๊ฐ€๋“ค์€ ์ด ์ด๋ฏธ์ง€๊ฐ€ ๊ธฐ์ˆ ์ด ์ธ๊ฐ„ ํฌ๋ฆฌ์—์ดํ„ฐ๋ฅผ ์ง“๋ฐŸ๋Š”๋‹ค๋Š” ์˜๋ฏธ๋กœ ํ•ด์„๋  ์ˆ˜ ์žˆ๋‹ค๊ณ  ์ฃผ์žฅํ•œ๋‹ค. ๋˜ํ•œ AI๋กœ ์ธํ•ด ํฌ๋ฆฌ์—์ดํ„ฐ๋“ค์ด ๋ฐ€๋ฆฌ๊ณ  ์žˆ๋‹ค๋Š” ์ƒํ™ฉ์„ ์—ฐ์ƒ์‹œํ‚จ๋‹ค๋Š” ์šฐ๋ ค์˜ ๋ชฉ์†Œ๋ฆฌ๋„ ๋‚˜์˜จ๋‹ค."

"KULLM3":"์ด ๋…ผ๋ž€์€ ์• ํ”Œ์ด ์œ ํŠœ๋ธŒ์— 7์ผ์— ๋ฐœํ‘œํ•œ ์ตœ์‹  iPad Pro ๊ด‘๊ณ  ์ฃผ์œ„์—์„œ ๋ฐœ์ƒํ–ˆ์Šต๋‹ˆ๋‹ค. ๊ด‘๊ณ ์—์„œ๋Š” ์•…๊ธฐ, ์กฐ๊ฐ์ƒ, ์นด๋ฉ”๋ผ, ๊ทธ๋ฆฌ๊ณ  ๋ฌผ๊ฐ์ด ์••์ถ•๊ธฐ์—์„œ ํŒŒ๊ดด๋˜๋Š” ๋ชจ์Šต์ด ๋ณด์—ฌ์ง€๊ณ , ๊ทธ ์ž๋ฆฌ์— iPad Pro๊ฐ€ ๋‚˜ํƒ€๋‚ฉ๋‹ˆ๋‹ค. ์ด๋Š” ์ƒˆ๋กœ์šด iPad Pro์˜ ์ธ๊ณต์ง€๋Šฅ ๊ธฐ๋Šฅ, ๊ณ ๊ธ‰ ๋””์Šคํ”Œ๋ ˆ์ด, ์„ฑ๋Šฅ, ๊ทธ๋ฆฌ๊ณ  ์–‡์€ ๋””์ž์ธ์„ ๊ฐ•์กฐํ•˜๋Š” ๊ฒƒ์œผ๋กœ ๋ณด์ž…๋‹ˆ๋‹ค. ์• ํ”Œ์€ ์ตœ์‹  'M4' ์นฉ์„ ํƒ‘์žฌํ•œ ์ƒˆ๋กœ ๊ณต๊ฐœ๋œ iPad Pro๊ฐ€ ์ž์‚ฌ ์—ญ์‚ฌ์ƒ ๊ฐ€์žฅ ์–‡์€ ๊ธฐ๊ธฐ๋ผ๊ณ  ์–ธ๊ธ‰ํ–ˆ์Šต๋‹ˆ๋‹ค. ํ•˜์ง€๋งŒ ๊ด‘๊ณ ๋Š” ์ถœ์‹œ ์งํ›„ ์ฆ‰๊ฐ์ ์ธ ๋ฐ˜๋ฐœ์„ ๋ฐ›์•˜์Šต๋‹ˆ๋‹ค. ๊ด‘๊ณ ์—์„œ๋Š” ์ฐฝ์ž‘์ž๋ฅผ ์ƒ์ง•ํ•˜๋Š” ๋ฌผ๊ฑด๋“ค์ด ํŒŒ๊ดด๋˜๋Š” ๋ชจ์Šต์ด ๊ทธ๋ž˜ํ”ฝํ•˜๊ฒŒ ๋ณด์—ฌ์ง€๊ธฐ ๋•Œ๋ฌธ์ž…๋‹ˆ๋‹ค. ๋น„ํŒ์ž๋“ค์€ ์ด ์ด๋ฏธ์ง€๊ฐ€ ๊ธฐ์ˆ ์ด ์ธ๊ฐ„ ์ฐฝ์ž‘์ž๋ฅผ ์••๋„ํ•˜๋Š” ๊ฒƒ์œผ๋กœ ํ•ด์„๋  ์ˆ˜ ์žˆ๋‹ค๊ณ  ์ฃผ์žฅํ•˜๋ฉฐ, ์ผ๋ถ€๋Š” ์ด๊ฐ€ ์ฐฝ์ž‘์ž๋“ค์ด AI ๋•Œ๋ฌธ์— ์ง€์œ„๋ฅผ ์žƒ๋Š” ์ƒํ™ฉ์„ ์—ฐ์ƒ์‹œํ‚จ๋‹ค๊ณ  ์šฐ๋ คํ•˜๊ณ  ์žˆ์Šต๋‹ˆ๋‹ค."
"EEVE-10.8b-it":ํ•ด๋‹น ๋…ผ๋ž€์€ ์• ํ”Œ์ด 7์ผ์— ์œ ํŠœ๋ธŒ์— ๊ณต๊ฐœํ•œ ์ตœ์‹  ์•„์ดํŒจ๋“œ ํ”„๋กœ ๊ด‘๊ณ ์™€ ๊ด€๋ จํ•˜์—ฌ ๋ฐœ์ƒํ–ˆ์Šต๋‹ˆ๋‹ค. ํ•ด๋‹น ๊ด‘๊ณ ์—์„œ๋Š” ์•…๊ธฐ, ์กฐ๊ฐ์ƒ, ์นด๋ฉ”๋ผ, ๊ทธ๋ฆฌ๊ณ  ๋ถ“์ด ๋ˆŒ๋Ÿฌ์ ธ ๋ถ€์„œ์ง€๋Š” ๋ชจ์Šต๊ณผ ํ•จ๊ป˜ ๊ทธ ์ž๋ฆฌ์— ์ƒˆ๋กœ์šด ์•„์ดํŒจ๋“œ ํ”„๋กœ๊ฐ€ ๋“ฑ์žฅํ•˜๋Š” ์žฅ๋ฉด์„ ์ƒ์ƒํ•˜๊ฒŒ ๋ณด์—ฌ์ฃผ๊ณ  ์žˆ์Šต๋‹ˆ๋‹ค. ์ด๋Š” ์ƒˆ๋กœ์šด ์•„์ดํŒจ๋“œ ํ”„๋กœ์˜ ์ธ๊ณต์ง€๋Šฅ ๊ธฐ๋Šฅ, ์ง„๋ณด๋œ ๋””์Šคํ”Œ๋ ˆ์ด, ์„ฑ๋Šฅ, ๊ทธ๋ฆฌ๊ณ  ๋‘๊ป˜๋ฅผ ๋ถ€๊ฐ์‹œํ‚ค๊ณ ์ž ํ•˜๋Š” ๊ฒƒ์œผ๋กœ ๋ณด์ž…๋‹ˆ๋‹ค. ์• ํ”Œ์€ ๊ฐ“ ๋ฐœํ‘œ๋œ ์•„์ดํŒจ๋“œ ํ”„๋กœ๊ฐ€ ์ตœ์‹  'M4' ์นฉ์„ ํƒ‘์žฌํ•˜๊ณ  ์žˆ์œผ๋ฉฐ, ์• ํ”Œ ์ œํ’ˆ ์ค‘ ๊ฐ€์žฅ ์–‡์€ ์žฅ์น˜๋ผ๊ณ  ์–ธ๊ธ‰ํ–ˆ์Šต๋‹ˆ๋‹ค. ๊ด‘๊ณ ๊ฐ€ ์ถœ์‹œ๋˜์ž๋งˆ์ž ๋ฐ”๋กœ ๋ญ‡๋งค๋ฅผ ๋งž์•˜๋Š”๋ฐ, ์ด๋Š” ์ฐฝ์ž‘์ž๋“ค์„ ์ƒ์ง•ํ•˜๋Š” ๋ฌผ๊ฑด๋“ค์ด ๋ˆŒ๋ ค ๋ถ€์„œ์ง€๋Š” ์ž”์ธํ•œ ์žฅ๋ฉด์„ ๋‹ด๊ณ  ์žˆ๊ธฐ ๋•Œ๋ฌธ์ž…๋‹ˆ๋‹ค. ๋น„ํŒ์ž๋“ค์€ ์ด๋Ÿฌํ•œ ์ด๋ฏธ์ง€๊ฐ€ ๊ธฐ์ˆ ์ด ์ธ๊ฐ„ ์ฐฝ์ž‘์ž๋“ค์„ ์ง“๋ฐŸ์€ ๊ฒƒ์œผ๋กœ ํ•ด์„๋  ์ˆ˜ ์žˆ๋‹ค๊ณ  ์ฃผ์žฅํ•˜๊ณ  ์žˆ์Šต๋‹ˆ๋‹ค. ์ผ๋ถ€ ์‚ฌ๋žŒ๋“ค์€ ๋˜ํ•œ AI๋กœ ์ธํ•ด ์ฐฝ์ž‘์ž๋“ค์ด ๋„ํƒœ๋˜๊ณ  ์žˆ๋Š” ์ƒํ™ฉ์„ ์•”์‹œํ•˜๋Š” ๊ฒƒ ๊ฐ™์•„ ์šฐ๋ ค๋ฅผ ํ‘œํ•˜๊ณ  ์žˆ์Šต๋‹ˆ๋‹ค."
"Seagull-13B":"์ด ๋…ผ๋ž€์ด ๋ฐœ์ƒํ•œ ๊ฒƒ์€ 7์ผ์— Apple์ด YouTube์— ๊ณต๊ฐœํ•œ ์ตœ์‹  iPad Pro์˜ ์ƒˆ๋กœ์šด ๊ด‘๊ณ ์™€ ๊ด€๋ จ์ด ์žˆ์Šต๋‹ˆ๋‹ค. ์ด ๊ด‘๊ณ ๋Š” ์•…๊ธฐ, ๋™์ƒ, ์นด๋ฉ”๋ผ, ๋ฌผ๊ฐ์„ ํ”„๋ ˆ์Šค์— ๋ˆ„๋ฅด๋Š” ์žฅ๋ฉด์„ ๋ณด์—ฌ์ค€ ๋‹ค์Œ ๊ทธ ์ž๋ฆฌ์— iPad Pro๊ฐ€ ๋‚˜ํƒ€๋‚ฉ๋‹ˆ๋‹ค. ์ƒˆ๋กœ์šด iPad Pro์˜ ์ธ๊ณต์ง€๋Šฅ ๊ธฐ๋Šฅ, ๊ณ ๊ธ‰ ๋””์Šคํ”Œ๋ ˆ์ด, ์„ฑ๋Šฅ ๋ฐ ๋‘๊ป˜๋ฅผ ๊ฐ•์กฐํ•˜๋Š” ๊ฒƒ์œผ๋กœ ๋ณด์ž…๋‹ˆ๋‹ค. Apple์€ ์ƒˆ๋กœ ๊ณต๊ฐœ ๋œ iPad Pro์— ์ตœ์‹  'M4' ์นฉ์ด ์žฅ์ฐฉ๋˜์–ด ์žˆ์œผ๋ฉฐ Apple ์—ญ์‚ฌ์ƒ ๊ฐ€์žฅ ์–‡์€ ์žฅ์น˜๋ผ๊ณ  ์–ธ๊ธ‰ํ–ˆ์Šต๋‹ˆ๋‹ค. ์ด ๊ด‘๊ณ ๋Š” ํฌ๋ฆฌ์—์ดํ„ฐ๋ฅผ ์ƒ์ง•ํ•˜๋Š” ๋ฌผ์ฒด๋ฅผ ๊ทธ๋ž˜ํ”ฝ์œผ๋กœ ํ‘œํ˜„ํ•˜์—ฌ ๊ณต๊ฐœ ์งํ›„ ๋ฐ˜๋ฐœ์„ ๋ถˆ๋Ÿฌ์ผ์œผ์ผฐ์Šต๋‹ˆ๋‹ค. ๋น„ํ‰๊ฐ€๋“ค์€ ์ด ์ด๋ฏธ์ง€๊ฐ€ ๋”ฅ๋Ÿฌ๋‹์ด ์ธ๊ฐ„ ์ฐฝ์ž‘์ž๋ฅผ ์ง“๋ฐŸ๋Š” ๊ธฐ์ˆ ๋กœ ํ•ด์„๋  ์ˆ˜ ์žˆ๋‹ค๊ณ  ์ง€์ ํ•ฉ๋‹ˆ๋‹ค. ์ผ๋ถ€๋Š” AI๋กœ ์ธํ•ด ํฌ๋ฆฌ์—์ดํ„ฐ๊ฐ€ ํ† ๋Œ€๋ฅผ ์žƒ๋Š” ์ƒํ™ฉ์„ ์—ฐ์ƒ์‹œํ‚จ๋‹ค๋Š” ์šฐ๋ ค์˜ ๋ชฉ์†Œ๋ฆฌ๋„ ์ œ๊ธฐํ–ˆ์Šต๋‹ˆ๋‹ค.\n"
"Synatra-7B":"์ด ๋…ผ๋ž€์€ 7์ผ Apple์ด YouTube์— ์˜ฌ๋ฆฐ ์ตœ์‹  iPad Pro์˜ ์‹ ๊ทœ ๊ด‘๊ณ ๋ฅผ ์ค‘์‹ฌ์œผ๋กœ ๋ฒŒ์–ด์กŒ์Šต๋‹ˆ๋‹ค. ์ด ๊ด‘๊ณ ์—๋Š” ๋ชฉ๊ด€์•…๊ธฐ, ์กฐ๊ฐ์ƒ, ์นด๋ฉ”๋ผ, ๋ฌผ๊ฐ์ด ์••๋ ฅ์„ ๊ฐ€ํ•œ ํ›„ ๊ทธ ์ž๋ฆฌ์— iPad Pro๊ฐ€ ๋‚˜ํƒ€๋‚˜๋Š” ์žฅ๋ฉด์ด ํ‘œ์‹œ๋˜๋Š” ๋ฌด๋‹จ์žฅ์‹ ๊ด‘๊ณ ์ž…๋‹ˆ๋‹ค. ๊ทธ ๊ด‘๊ณ ๋Š” ์ƒˆ iPad Pro์˜ ์ธ๊ณต์ง€๋Šฅ ๊ธฐ๋Šฅ, ๊ณ ๊ธ‰ ๋””์Šคํ”Œ๋ ˆ์ด, ์„ฑ๋Šฅ ๋ฐ ๋‘๊ป˜๋ฅผ ๊ฐ•์กฐํ•˜๋Š” ๊ฒƒ์œผ๋กœ ๋ณด์ž…๋‹ˆ๋‹ค. Apple์€ ์ƒˆ๋กญ๊ฒŒ ๋ฐœํ‘œ๋œ iPad Pro์—๋Š” ์ตœ์‹  'M4' ์นฉ์ด ํƒ‘์žฌ๋˜์–ด ์žˆ์œผ๋ฉฐ ์—ญ์‚ฌ์ƒ ๊ฐ€์žฅ ์–‡์€ ๊ธฐ๊ธฐ๋ผ๊ณ  ์–ธ๊ธ‰ํ–ˆ์Šต๋‹ˆ๋‹ค. ์ด ๊ด‘๊ณ ๋Š” ํฌ๋ฆฌ์—์ดํ„ฐ๋ฅผ ์ƒ์ง•ํ•˜๋Š” ๋ฌผ์ฒด๊ฐ€ ์ง“๊ธฐ์— ๋งž์„œ ์žˆ๋‹ค๋Š” ๋ชจ์Šต์„ ๊ทธ๋ž˜ํ”ฝ์œผ๋กœ ํ‘œํ˜„ํ•œ ํ›„ ์ฆ‰์‹œ ๋ฐ˜๋ฐœ์„ ๋ถˆ๋Ÿฌ ์ผ์œผ์ผฐ์Šต๋‹ˆ๋‹ค. ๋น„ํ‰๊ฐ€๋“ค์€ ์ด ์ด๋ฏธ์ง€๊ฐ€ ๊ธฐ์ˆ ์ด ์ธ๊ฐ„ ํฌ๋ฆฌ์—์ดํ„ฐ๋ฅผ ํ˜‘๋ฐ•ํ•˜๋Š” ๊ฒƒ์œผ๋กœ ํ•ด์„๋  ์ˆ˜ ์žˆ๋‹ค๊ณ  ์ฃผ์žฅํ•ฉ๋‹ˆ๋‹ค. ์ผ๋ถ€๋Š” ๋˜ํ•œ AI๋กœ ์ธํ•ด ํฌ๋ฆฌ์—์ดํ„ฐ๊ฐ€ ์ง€์œ„๋ฅผ ์žƒ๋Š” ์ƒํ™ฉ์„ ๋ถˆ๋Ÿฌ์ผ์œผํ‚ฌ ์ˆ˜ ์žˆ๋‹ค๊ณ  ์šฐ๋ คํ•˜๋Š” ๋ชฉ์†Œ๋ฆฌ๋„ ์žˆ์Šต๋‹ˆ๋‹ค."
"nhndq-nllb":"์ด ๋…ผ๋ž€์€ ์• ํ”Œ์ด 7์ผ ์œ ํŠœ๋ธŒ์— ๊ณต๊ฐœํ•œ ์ตœ์‹  ์•„์ดํŒจ๋“œ ํ”„๋กœ์˜ ์ƒˆ ๊ด‘๊ณ ๋ฅผ ๋‘˜๋Ÿฌ์‹ธ๊ณ  ๋ถˆ๊ฑฐ์กŒ๋‹ค. ๊ด‘๊ณ ๋Š” ์•…๊ธฐ, ์กฐ๊ฐ์ƒ, ์นด๋ฉ”๋ผ, ํŽ˜์ธํŠธ ๋“ฑ์ด ํ”„๋ ˆ์Šค์—์„œ ์œผ๊นจ์ง€๊ณ  ๊ทธ ์ž๋ฆฌ์— ์•„์ดํŒจ๋“œ ํ”„๋กœ๊ฐ€ ๋“ฑ์žฅํ•˜๋Š” ๋ชจ์Šต์„ ๋ณด์—ฌ์ค€๋‹ค. ์ด๋Š” ์ƒˆ๋กœ์šด ์•„์ดํŒจ๋“œ ํ”„๋กœ์˜ ์ธ๊ณต์ง€๋Šฅ ๊ธฐ๋Šฅ๊ณผ ๊ณ ๊ธ‰ ๋””์Šคํ”Œ๋ ˆ์ด, ์„ฑ๋Šฅ, ๋‘๊ป˜ ๋“ฑ์„ ๊ฐ•์กฐํ•˜๋Š” ๊ฒƒ์œผ๋กœ ๋ณด์ธ๋‹ค. ์• ํ”Œ์€ ์ƒˆ๋กœ ๊ณต๊ฐœ๋œ ์•„์ดํŒจ๋“œ ํ”„๋กœ๊ฐ€ ์ตœ์‹  'M4' ์นฉ์„ ์žฅ์ฐฉํ•˜๊ณ  ์žˆ์œผ๋ฉฐ ์• ํ”Œ ์—ญ์‚ฌ์ƒ ๊ฐ€์žฅ ์–‡์€ ์žฅ์น˜๋ผ๊ณ  ์–ธ๊ธ‰ํ–ˆ๋‹ค. AI๋กœ ์ธํ•ด ์ฆ‰๊ฐ"

"our-tech":"์ด๋ฒˆ ๋…ผ๋ž€์€ ์• ํ”Œ์ด ์ง€๋‚œ 7์ผ ์œ ํŠœ๋ธŒ์— ๊ณต๊ฐœํ•œ ์ตœ์‹  ์•„์ดํŒจ๋“œ ํ”„๋กœ ๊ด‘๊ณ ๋ฅผ ๋‘˜๋Ÿฌ์‹ธ๊ณ  ๋ถˆ๊ฑฐ์กŒ๋‹ค. ๊ด‘๊ณ ๋Š” ์•…๊ธฐ, ์กฐ๊ฐ์ƒ, ์นด๋ฉ”๋ผ, ๋ฌผ๊ฐ ๋“ฑ์„ ์••์ฐฉ๊ธฐ์— ๋„ฃ์–ด ๋ถ€์ˆด๋ฒ„๋ฆฌ๋‹ค๊ฐ€ ๊ทธ ์ž๋ฆฌ์— ์•„์ดํŒจ๋“œ ํ”„๋กœ๊ฐ€ ๋“ฑ์žฅํ•˜๋Š” ๊ฒƒ์œผ๋กœ, ์ƒˆ๋กœ ๊ณต๊ฐœ๋œ ์•„์ดํŒจ๋“œ ํ”„๋กœ๊ฐ€ ์ตœ์‹  'M4'์นฉ์„ ํƒ‘์žฌํ•˜๊ณ  ์• ํ”Œ ์‚ฌ์ƒ ๊ฐ€์žฅ ์–‡์€ ๊ธฐ๊ธฐ๋ผ๋Š” ์ ์„ ๊ฐ•์กฐํ•œ ๊ฒƒ์œผ๋กœ ๋ณด์ธ๋‹ค. ๊ด‘๊ณ ๋Š” ์ถœ์‹œ ์ฆ‰์‹œ ์ฐฝ์ž‘์ž๋ฅผ ์ƒ์ง•ํ•˜๋Š” ๋ฌผ๊ฑด๋“ค์ด ์••์ฐฉ๊ธฐ์— ๊ฐˆ๊ฒจ๋ฒ„๋ฆฌ๋Š” ์žฅ๋ฉด์„ ๊ทธ๋ž˜ํ”ฝ์œผ๋กœ ๋ณด์—ฌ์ค˜, ๊ธฐ์ˆ ์ด ์ธ๊ฐ„ ์ฐฝ์ž‘์ž๋ฅผ ์ง“๋ฐŸ๋Š” ๊ฒƒ์œผ๋กœ ํ•ด์„๋  ์ˆ˜ ์žˆ๋‹ค๋Š” ์ง€์ ๊ณผ ํ•จ๊ป˜, AI๋กœ ์ธํ•ด ์ฐฝ์ž‘์ž๋“ค์ด ์ง€์œ„๋ฅผ ์žƒ์–ด๊ฐ€๋Š” ์ƒํ™ฉ์„ ์—ฐ์ƒ์‹œํ‚จ๋‹ค๋Š” ๋น„ํŒ์ด ์ œ๊ธฐ๋๋‹ค."
"our-general":์ด๋ฒˆ ๋…ผ๋ž€์€ ์• ํ”Œ์ด ์ง€๋‚œ 7์ผ ์œ ํŠœ๋ธŒ์— ๊ณต๊ฐœํ•œ ์ตœ์‹  ์•„์ดํŒจ๋“œ ํ”„๋กœ ๊ด‘๊ณ ๋ฅผ ์ค‘์‹ฌ์œผ๋กœ ๋ถˆ๊ฑฐ์กŒ๋‹ค. ๊ด‘๊ณ ๋Š” ์•…๊ธฐ, ์กฐ๊ฐ์ƒ, ์นด๋ฉ”๋ผ, ๋ฌผ๊ฐ ๋“ฑ์„ ๋ˆ„๋ฅด๊ธฐ์— ์ถฉ๋ถ„ํ•œ ํž˜์„ ๊ฐ€์ง„ ํ”„๋ ˆ์Šค์— ์ง‘์–ด๋„ฃ๊ณ  ์œผ๊นจ๋Š” ๋ชจ์Šต์„ ๋ณด์—ฌ์ค€๋‹ค. ์ด์–ด ๊ทธ ์ž๋ฆฌ์— ์•„์ดํŒจ๋“œ ํ”„๋กœ๊ฐ€ ๋“ฑ์žฅํ•˜๋Š” ๊ฒƒ์œผ๋กœ, ์ƒˆ๋กœ ๊ณต๊ฐœ๋œ ์•„์ดํŒจ๋“œ ํ”„๋กœ๊ฐ€ ์ตœ์‹  'M4' ์นฉ์„ ํƒ‘์žฌํ•˜๊ณ  ์• ํ”Œ ์—ญ์‚ฌ์ƒ ๊ฐ€์žฅ ์–‡์€ ๊ธฐ๊ธฐ๋ผ๋Š” ์ ์„ ๊ฐ•์กฐํ•œ ๊ฒƒ์œผ๋กœ ๋ณด์ธ๋‹ค. ์ด ๊ด‘๊ณ ๋Š” ๊ณต๊ฐœ ์งํ›„๋ถ€ํ„ฐ ๋…ผ๋ž€์ด ์ผ์—ˆ๋Š”๋ฐ, ์ฐฝ์ž‘์ž๋ฅผ ์ƒ์ง•ํ•˜๋Š” ๋ฌผ๊ฑด๋“ค์ด ์œผ๊นจ์ง€๋Š” ์žฅ๋ฉด์ด ๊ทธ๋Œ€๋กœ ๋‹ด๊ฒจ์žˆ์–ด ๊ธฐ์ˆ ์ด ์ฐฝ์ž‘์ž๋ฅผ ์ง“๋ฐŸ๋Š”๋‹ค๋Š” ํ•ด์„์ด ๋‚˜์˜ฌ ์ˆ˜ ์žˆ๋‹ค๋Š” ์ง€์ ์ด ๋‚˜์™”๋‹ค. ๋˜ AI์— ๋ฐ€๋ ค ์ฐฝ์ž‘์ž๋“ค์ด ํž˜์„ ์žƒ๋Š” ์ƒํ™ฉ์„ ์—ฐ์ƒ์‹œํ‚จ๋‹ค๋Š” ์šฐ๋ ค๋„ ์ œ๊ธฐ๋๋‹ค."
"our-sharegpt":"7์ผ, Apple์ด YouTube์— ๊ณต๊ฐœํ•œ ์ตœ์‹  iPad Pro์˜ ์ƒˆ๋กœ์šด ๊ด‘๊ณ ์™€ ๊ด€๋ จํ•˜์—ฌ ๋…ผ๋ž€์ด ์ผ์–ด๋‚ฌ์Šต๋‹ˆ๋‹ค. ์ด ๊ด‘๊ณ ๋Š” ์•…๊ธฐ, ์กฐ๊ฐ์ƒ, ์นด๋ฉ”๋ผ, ๋ฌผ๊ฐ์ด ํ”„๋ ˆ์Šค์—์„œ ๋ถ€์„œ์ง€๋Š” ์žฅ๋ฉด์„ ๋ณด์—ฌ์ค€ ํ›„ ๊ทธ ์ž๋ฆฌ์— iPad Pro๊ฐ€ ๋“ฑ์žฅํ•ฉ๋‹ˆ๋‹ค. ์ƒˆ๋กœ์šด iPad Pro์˜ ์ธ๊ณต ์ง€๋Šฅ ๊ธฐ๋Šฅ, ๊ณ ๊ธ‰ ๋””์Šคํ”Œ๋ ˆ์ด, ์„ฑ๋Šฅ ๋ฐ ๋‘๊ป˜๋ฅผ ๊ฐ•์กฐํ•˜๋Š” ๊ฒƒ์œผ๋กœ ๋ณด์ž…๋‹ˆ๋‹ค. Apple์€ ์ƒˆ๋กœ ๊ณต๊ฐœ๋œ iPad Pro๊ฐ€ ์ตœ์‹  'M4' ์นฉ์ด ํƒ‘์žฌ๋˜์–ด ์žˆ์œผ๋ฉฐ Apple ์—ญ์‚ฌ์ƒ ๊ฐ€์žฅ ์–‡์€ ๊ธฐ๊ธฐ๋ผ๊ณ  ์–ธ๊ธ‰ํ–ˆ์Šต๋‹ˆ๋‹ค. ์ด ๊ด‘๊ณ ๋Š” ํฌ๋ฆฌ์—์ดํ„ฐ๋ฅผ ์ƒ์ง•ํ•˜๋Š” ๋ฌผ์ฒด๊ฐ€ ๋ถ€์„œ์ง€๋Š” ๊ฒƒ์„ ๊ทธ๋ž˜ํ”ฝ์œผ๋กœ ๋ฌ˜์‚ฌํ•˜๊ณ  ์žˆ์–ด ์ถœ์‹œ์™€ ๋™์‹œ์— ๋ฐ˜๋ฐœ์„ ๋ถˆ๋Ÿฌ์ผ์œผ์ผฐ์Šต๋‹ˆ๋‹ค. ๋น„ํ‰๊ฐ€๋“ค์€ ์ด ์ด๋ฏธ์ง€๊ฐ€ ๊ธฐ์ˆ ์ด ์ธ๊ฐ„ ํฌ๋ฆฌ์—์ดํ„ฐ๋ฅผ ์ง“๋ฐŸ๋Š” ๊ฒƒ์œผ๋กœ ํ•ด์„๋  ์ˆ˜ ์žˆ๋‹ค๊ณ  ์ฃผ์žฅํ•ฉ๋‹ˆ๋‹ค. ๋˜ํ•œ ์ผ๋ถ€์—์„œ๋Š” ํฌ๋ฆฌ์—์ดํ„ฐ๊ฐ€ ์ธ๊ณต์ง€๋Šฅ์œผ๋กœ ์ธํ•ด ์ฃผ๋ˆ… ๋“ค๊ณ  ์žˆ๋Š” ์ƒํ™ฉ์„ ์—ฐ์ƒ์‹œํ‚จ๋‹ค๊ณ  ์šฐ๋ คํ•˜๋Š” ๋ชฉ์†Œ๋ฆฌ๋„ ์žˆ์Šต๋‹ˆ๋‹ค."
```

<br><br>

# **Evalution Result**
์˜์–ด->ํ•œ๊ตญ์–ด ๋ฒˆ์—ญ ์„ฑ๋Šฅ์„ ํ‰๊ฐ€ํ•˜๊ธฐ์œ„ํ•œ ๋ฐ์ดํ„ฐ์…‹์„ ์„ ์ •ํ•˜์—ฌ ํ‰๊ฐ€๋ฅผ ์ง„ํ–‰ํ•˜์˜€์Šต๋‹ˆ๋‹ค.

### **ํ‰๊ฐ€ ๋ฐ์ดํ„ฐ์…‹ ์ถœ์ฒ˜**
- Aihub/FLoRes: [traintogpb/aihub-flores-koen-integrated-sparta-30k](https://huggingface.co/datasets/traintogpb/aihub-flores-koen-integrated-sparta-30k) | (test set 1k)
- iwslt-2023  : [shreevigneshs/iwslt-2023-en-ko-train-val-split-0.1](https://huggingface.co/datasets/shreevigneshs/iwslt-2023-en-ko-train-val-split-0.1) | (f_test 597, if_test 597)
- ko_news_2024: [nayohan/ko_news_eval40](https://huggingface.co/datasets/nayohan/ko_news_eval40) | (40)

### **๋ชจ๋ธ ํ‰๊ฐ€๋ฐฉ๋ฒ•**
- ๊ฐ ๋ชจ๋ธ์€ ํ—ˆ๊น…ํŽ˜์ด์Šค์— ReadMe์— ์ ํ˜€์žˆ๋Š” ์ถ”๋ก ์ฝ”๋“œ๋ฅผ ๊ธฐ๋ฐ˜์œผ๋กœ ๊ฐ๊ฐ ์ถ”๋ก ํ•˜์˜€์Šต๋‹ˆ๋‹ค. (๊ณตํ†ต: max_new_tokens=512) 
- EEVE๋Š” ๋ช…๋ น์–ด("๋‹น์‹ ์€ ๋ฒˆ์—ญ๊ธฐ ์ž…๋‹ˆ๋‹ค. ์˜์–ด๋ฅผ ํ•œ๊ตญ์–ด๋กœ ๋ฒˆ์—ญํ•˜์„ธ์š”.")๋ฅผ ์‹œ์Šคํ…œํ”„๋กฌํ”„ํŠธ์— ์ถ”๊ฐ€ํ•˜์˜€๊ณ , KULLM3๋Š” ๊ธฐ์กด ์‹œ์Šคํ…œํ”„๋กฌํ”„ํŠธ๋ฅผ ์œ ์ง€ํ•˜๊ณ , ์œ ์ €์˜ ์ž…๋ ฅ ๋งจ ์•ž์— ์ถ”๊ฐ€ํ•˜์˜€์Šต๋‹ˆ๋‹ค.

<br>

## **Aihub ์˜-ํ•œ ๋ฒˆ์—ญ๋ฐ์ดํ„ฐ์…‹ ํ‰๊ฐ€**
* [Aihub ํ‰๊ฐ€ ๋ฐ์ดํ„ฐ์…‹](https://huggingface.co/datasets/traintogpb/aihub-flores-koen-integrated-sparta-30k)์€ ๋ชจ๋ธ๋“ค์ด ํ•™์Šต๋ฐ์ดํ„ฐ์…‹์— ํฌํ•จ๋˜์—ˆ์„ ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค. ์นดํ…Œ๊ณ ๋ฆฌ๋ณ„ ์„ฑ๋Šฅ์„ ํ™•์ธํ•˜๋Š” ์šฉ๋„๋กœ๋งŒ ์ฐธ๊ณ ํ•ด์ฃผ์„ธ์š”. [[์นดํ…Œ๊ณ ๋ฆฌ ์„ค๋ช… ๋งํฌ]](https://huggingface.co/datasets/traintogpb/aihub-koen-translation-integrated-tiny-100k)
![image/png](https://cdn-uploads.huggingface.co/production/uploads/6152b4b9ecf3ca6ab820e325/TMo05LOUhPGYNbT2ADOgi.png)
| model            |   aihub-111 |   aihub-124 |   aihub-125 |   aihub-126 |   aihub-563 |   aihub-71265 |   aihub-71266 |   aihub-71382 |   average |
|:-----------------|------------:|------------:|------------:|------------:|------------:|--------------:|--------------:|--------------:|----------:|
| [EEVE-10.8b-it](https://huggingface.co/yanolja/EEVE-Korean-10.8B-v1.0)    |        6.15 |       11.81 |        5.78 |        4.99 |        6.31 |         10.99 |          9.41 |          6.44 |    7.73   |
| [KULLM3](https://huggingface.co/nlpai-lab/KULLM3)           |        9.00 |       13.49 |       10.43 |        5.90 |        1.92 |         16.37 |         10.02 |          8.39 |    9.44   |
| [Seagull-13B](https://huggingface.co/kuotient/Seagull-13b-translation)      |        9.8  |       18.38 |        8.51 |        5.53 |        8.74 |         17.44 |         10.11 |         11.21 |   11.21   |
| [Synatra-7B](https://huggingface.co/maywell/Synatra-7B-v0.3-Translation)       |        6.99 |       25.14 |        7.79 |        5.31 |        9.95 |         19.27 |         13.20 |          8.93 |   12.07   |
| [nhndq-nllb](https://huggingface.co/NHNDQ/nllb-finetuned-en2ko)       |       24.09 |       48.71 |       22.89 |       13.98 |       18.71 |         30.18 |         32.49 |         18.62 |   26.20   |
| [our-tech](nayohan/llama3-8b-it-translation-tech-en-ko-1sent)         |       20.19 |       37.48 |       18.50 |       12.45 |       16.96 |         13.92 |         43.54 |          9.62 |   21.58   |
| [our-general](https://huggingface.co/nayohan/llama3-8b-it-translation-general-en-ko-1sent)      |       24.72 |       45.22 |       21.61 |       18.97 |       17.23 |         30.00 |         32.08 |         13.55 |   25.42   |
| [our-sharegpt](https://huggingface.co/nayohan/llama3-8b-it-translation-sharegpt-en-ko)     |       12.42 |       19.23 |       10.91 |        9.18 |       14.30 |         26.43 |         12.62 |         15.57 |   15.08   |
| **our-instrucTrans** |       24.89 |       47.00 |       22.78 |       21.78 |       24.27 |         27.98 |         31.31 |         15.42 |**26.92**  |
## **FLoRes ์˜-ํ•œ ๋ฒˆ์—ญ๋ฐ์ดํ„ฐ์…‹ ํ‰๊ฐ€**
[FloRes](https://huggingface.co/datasets/facebook/flores)๋Š” ํŽ˜์ด์Šค๋ถ์—์„œ ๊ณต๊ฐœํ•œ ์˜์–ด์™€ ์ ์€ ๋ฆฌ์†Œ์Šค์˜ ์–ธ์–ด 200๊ฐœ์— ๋Œ€ํ•ด์„œ ๋ณ‘๋ ฌ๋กœ ๊ตฌ์„ฑํ•œ ๋ฒˆ์—ญ ๋ฒค์น˜๋งˆํฌ ๋ฐ์ดํ„ฐ์…‹์ž…๋‹ˆ๋‹ค. 
[traintogpb/aihub-flores-koen-integrated-sparta-30k](https://huggingface.co/datasets/traintogpb/aihub-flores-koen-integrated-sparta-30k)๋ฅผ ํ™œ์šฉํ•˜์—ฌ ํ‰๊ฐ€๋ฅผ ์ง„ํ–‰ํ•˜์˜€์Šต๋‹ˆ๋‹ค. (ํ•œ๋ฌธ์žฅ ๊ตฌ์„ฑ)
![image/png](https://cdn-uploads.huggingface.co/production/uploads/6152b4b9ecf3ca6ab820e325/ZDeA-7e-0xfXaGOmyS9zs.png)
| model            |   flores-dev |   flores-devtest |   average |
|:-----------------|-------------:|-----------------:|----------:|
| EEVE-10.8b-it    |        10.99 |            11.71 |    11.35  |
| KULLM3           |        12.83 |            13.23 |    13.03  |
| Seagull-13B      |        11.48 |            11.99 |    11.73  |
| Synatra-7B       |        10.98 |            10.81 |    10.89  |
| nhndq-nllb       |        12.79 |            15.15 |    13.97  |
| our-tech         |        12.14 |            12.04 |    12.09  |
| our-general      |        14.93 |            14.58 |    14.75  |
| our-sharegpt     |        14.71 |            16.69 |    15.70  |
| our-instrucTrans |        14.49 |            17.69 | **16.09** |
## **iwslt-2023**
[iwslt-2023 ๋ฐ์ดํ„ฐ์…‹](https://huggingface.co/datasets/shreevigneshs/iwslt-2023-en-ko-train-val-split-0.1)์€ ๋™์ผํ•œ ์˜์–ด๋ฌธ์žฅ์„ ๊ฐ๊ฐ ๋ฐ˜๋ง, ์กด๋Œ“๋ง์˜ ํ•œ๊ตญ์–ด๋กœ ํ‰๊ฐ€๋ฐ์ดํ„ฐ์…‹์ด ๊ตฌ์„ฑ๋˜์–ด ์žˆ์Šต๋‹ˆ๋‹ค. ๋ชจ๋ธ์˜ ์กด๋Œ€/๋ฐ˜๋ง ๊ฒฝํ–ฅ์„ ์ƒ๋Œ€์ ์œผ๋กœ ํ™•์ธํ•  ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค. (ํ•œ๋ฌธ์žฅ ๊ตฌ์„ฑ)
![image/png](https://cdn-uploads.huggingface.co/production/uploads/6152b4b9ecf3ca6ab820e325/UJvuCnbjWokBWQNhD4L63.png)
| model            |         iwslt_zondae |      iwslt_banmal |   average |
|:-----------------|---------------------:|------------------:|----------:|
| EEVE-10.8b-it    |                 4.62 |              3.79 |     4.20  |
| KULLM3           |                 5.94 |              5.24 |     5.59  |
| Seagull-13B      |                 6.14 |              4.54 |     5.34  |
| Synatra-7B       |                 5.43 |              4.73 |     5.08  |
| nhndq-nllb       |                 8.36 |              7.44 |  **7.90** |
| our-tech         |                 3.99 |              3.95 |     3.97  |
| our-general      |                 7.33 |              6.18 |     6.75  |
| our-sharegpt     |                 7.83 |              6.35 |     7.09  |
| our-instrucTrans |                 8.63 |              6.97 |     7.80  |
## **ko_news_eval40**
[ko_news_eval40 ๋ฐ์ดํ„ฐ์…‹](https://huggingface.co/datasets/nayohan/ko_news_eval40)์€ ํ•™์Šต๋˜์ง€ ์•Š์•˜์„ ์ƒˆ๋กœ์šด ๋ฐ์ดํ„ฐ์…‹์— ํ‰๊ฐ€ํ•˜๊ณ ์ž 24๋…„5์›” ๋‰ด์Šค๋ฅผ ๊ฐ ์นดํ…Œ๊ณ ๋ฆฌ(4) ๋ณ„ 10๊ฐœ์”ฉ ๊ธฐ์‚ฌ ๋‚ด ๋ฌธ๋‹จ ์ผ๋ถ€๋ฅผ ์ˆ˜์ง‘ํ•˜๊ณ , GPT4๋กœ ๋ฒˆ์—ญํ•˜์—ฌ ๊ตฌ์„ฑํ•˜์˜€์Šต๋‹ˆ๋‹ค.
์˜์–ด๋ฅผ ์ผ์ƒ๋‰ด์Šค์— ์‚ฌ์šฉ๋˜๋Š” ํ•œ๊ตญ์–ด๋กœ ์ž˜ ๋ฒˆ์—ญํ•˜๋Š”์ง€๋ฅผ ํ‰๊ฐ€ํ•ฉ๋‹ˆ๋‹ค. (๋ฌธ๋‹จ ๊ตฌ์„ฑ)
![image/png](https://cdn-uploads.huggingface.co/production/uploads/6152b4b9ecf3ca6ab820e325/OaE5z_yQT9sIIz0zsn644.png)
| model            |   IT/๊ณผํ•™ |   ๊ฒฝ์ œ |   ์‚ฌํšŒ |   ์˜คํ”ผ๋‹ˆ์–ธ  |   average |
|:-----------------|----------:|-------:|-------:|------------:|----------:|
| EEVE-10.8b-it    |      9.03 |   6.42 |   5.56 |        5.10 |    6.52   |
| KULLM3           |      9.82 |   5.26 |   3.48 |        7.48 |    6.51   |
| Seagull-13B      |      7.41 |   6.78 |   4.76 |        4.85 |    5.95   |
| Synatra-7B       |     11.44 |   5.59 |   4.57 |        6.31 |    6.97   |
| nhndq-nllb       |     11.97 |  11.12 |   6.14 |        5.28 |    8.62   |
| our-tech         |     10.45 |   9.98 |   5.13 |       10.15 |    8.92   |
| our-general      |     16.22 |  10.61 |   8.51 |        7.33 |   10.66   |
| our-sharegpt     |     12.71 |   8.06 |   7.70 |        6.43 |    8.72   |
| our-instrucTrans |     20.42 |  12.77 |  11.40 |       10.31 |**13.72**  |
## **Average**
![image/png](https://cdn-uploads.huggingface.co/production/uploads/6152b4b9ecf3ca6ab820e325/bf2qjeg-03WRVTIbqvG7C.png)
| model            |   aihub |   flores |   iwslt |    news |   average |
|:-----------------|--------:|---------:|--------:|--------:|----------:|
| [EEVE-10.8b-it](https://huggingface.co/yanolja/EEVE-Korean-10.8B-v1.0)    |  7.73   |   11.35  |   4.20  |  6.52   |   7.45    |
| [KULLM3](https://huggingface.co/nlpai-lab/KULLM3)           |  9.44   |   13.03  |   5.59  |  6.51   |   8.64    |
| [Seagull-13B](https://huggingface.co/kuotient/Seagull-13b-translation)      | 11.21   |   11.73  |   5.34  |  5.95   |   8.56    |
| [Synatra-7B](https://huggingface.co/maywell/Synatra-7B-v0.3-Translation)       | 12.07   |   10.89  |   5.08  |  6.97   |   8.75    |
| [nhndq-nllb](https://huggingface.co/NHNDQ/nllb-finetuned-en2ko)       | 26.20   |   13.97  |**7.90** |  8.62   |  14.17    |
| [our-tech](nayohan/llama3-8b-it-translation-tech-en-ko-1sent)         | 21.58   |   12.09  |   3.97  |  8.92   |  11.64    |
| [our-general](https://huggingface.co/nayohan/llama3-8b-it-translation-general-en-ko-1sent)      | 25.42   |   14.75  |   6.75  | 10.66   |  14.40    |
| [our-sharegpt](https://huggingface.co/nayohan/llama3-8b-it-translation-sharegpt-en-ko)     | 15.08   |   15.70  |   7.09  |  8.72   |  11.64    |
| **our-instrucTrans** |**26.92**| **16.09**|   7.80  |**13.72**| **16.13** |
### **Citation**
```bibtex
@article{InstrcTrans8b,
  title={llama3-instrucTrans-enko-8b},
  author={Na, Yohan},
  year={2024},
  url={https://huggingface.co/nayohan/llama3-instrucTrans-enko-8b}
}
```
```bibtex
@article{llama3modelcard,
        title={Llama 3 Model Card},
        author={AI@Meta},
        year={2024},
        url={https://github.com/meta-llama/llama3/blob/main/MODEL_CARD.md}
}
```