Update README.md
Browse files
README.md
CHANGED
@@ -25,11 +25,21 @@ It achieves the following results on the evaluation set:
|
|
25 |
|
26 |
## Model description
|
27 |
|
28 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
29 |
|
30 |
## Intended uses & limitations
|
31 |
|
32 |
-
|
|
|
33 |
|
34 |
## Training and evaluation data
|
35 |
|
|
|
25 |
|
26 |
## Model description
|
27 |
|
28 |
+
ํ๋ก์ ํธ ์ฉ๋๋ก ํ์ธํ๋๋ ๋ชจ๋ธ์
๋๋ค.
|
29 |
+
OpenAI์ Whisper-Base ๋ชจ๋ธ์ ๋ฐํ์ผ๋ก 'ํ๊ตญ์ด ์ ์์ง ์์ฑ ํตํ ๋ฐ์ดํฐ'์ ๋ํ ์ ํ๋๋ฅผ ์ฆ๊ฐ์ํค๊ณ ์ ํ์ธํ๋์ ์งํํ ๋ชจ๋ธ์ด๋ฉฐ,
|
30 |
+
์ฌ์ฉํ ๋ฐ์ดํฐ๋ AI-HUB์ โ์ ์์ง ์ ํ๋ง ์์ฑ์ธ์ ๋ฐ์ดํฐโ ์ค ์ผ๋ถ๋ก์ ์ค๋์ค ํ์ผ ๊ธฐ์ค 240,771.06์ด(ํ์ผ 1๊ฐ๋น ํ๊ท ๊ธธ์ด๋ ์ฝ 5.296์ด)
|
31 |
+
ํ
์คํธ ๋ฐ์ดํฐ ๊ธฐ์ค ์ด 1,696,414๊ธ์์ ํฌ๊ธฐ์
๋๋ค.
|
32 |
+
|
33 |
+
This is a fine-tuned model for project use.
|
34 |
+
This model was fine-tuned to increase the accuracy of โKorean low-quality voice call dataโ based on OpenAIโs Whisper-Base model.
|
35 |
+
The data used is part of AI-HUBโs โlow-quality telephone network voice recognition dataโ,
|
36 |
+
which is 240,771.06 seconds based on audio files(average length per file is about 5.296 seconds).
|
37 |
+
The total size is 1,696,414 characters based on text data.
|
38 |
|
39 |
## Intended uses & limitations
|
40 |
|
41 |
+
ํ์ธํ๋์ ์ฌ์ฉ๋ Base model๊ณผ dataset ๋ชจ๋ ํ์ต ๋ชฉ์ ์ผ๋ก ์ฌ์ฉํ์์ผ๋ฉฐ,
|
42 |
+
๋ฐ๋ผ์ ๋ณธ ๋ชจ๋ธ ์ญ์ ํ์ต ๋ชฉ์ ์ผ๋ก๋ง ์ฌ์ฉ ๊ฐ๋ฅํฉ๋๋ค.
|
43 |
|
44 |
## Training and evaluation data
|
45 |
|