Hugging Face's logo

RecCode
/
whisper_final

Automatic Speech Recognition
Transformers
TensorBoard
Safetensors
whisper
Generated from Trainer
Model card Files
xet
 Metrics Community
RecCode commited on
Commit
030cfe1
Β·
verified Β·
1 Parent(s): 46dc872

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +16 -17
README.md CHANGED
@@ -15,29 +15,28 @@ should probably proofread and complete it, then remove this comment. -->
15
 
16
  # ꡬ음μž₯μ•  ν™˜μžλ₯Ό μœ„ν•œ μŒμ„±μΈμ‹ λͺ¨λΈ
17
 
18
- ## 곡λͺ¨μ „
19
- - μž¬λ‹¨λ²•μΈ λ―Έλž˜μ™€ μ†Œν”„νŠΈμ›¨μ–΄μ™€ ν•¨κ»˜ν•˜λŠ” 제 3νšŒμ•„μ΄λ””μ–΄ 곡λͺ¨μ „
20
 
21
- ## ν”„λ‘œμ νŠΈ
22
- - "ꡬ음μž₯μ•  μŒμ„± 데이터λ₯Ό ν™œμš©ν•œ κ³ λ Ή ν™˜μžμ˜ μ˜μ‚¬μ†Œν†΅ κ°œμ„ λ°©μ•ˆ"
23
 
24
- ## Model description
 
 
 
25
 
26
- λ³Έ λͺ¨λΈμ€ "ꡬ음μž₯μ•  μŒμ„± 데이터λ₯Ό ν™œμš©ν•œ κ³ λ Ή ν™˜μžμ˜ μ˜μ‚¬μ†Œν†΅ κ°œμ„ λ°©μ•ˆ" ν”„λ‘œμ νŠΈμ˜ ꡬ음μž₯μ• ν™˜μžμ— λŒ€ν•œ ν•œκ΅­μ–΄ μŒμ„±μΈμ‹ λͺ¨λΈμž„. OpenAI의 Whisper λͺ¨λΈμ„ νŒŒμΈνŠœλ‹ ν•˜μ—¬ ꡬ음μž₯μ• μ˜ μŒμ„±μ  νŠΉμ„±μ„ λ°˜μ˜ν•œ λͺ¨λΈμ„ κ΅¬μΆ•ν•˜μ˜€μŒ.
 
 
27
 
28
- ## Intended uses & limitations
29
 
30
- More information needed
 
31
 
32
- ## Training and evaluation data
33
 
34
- More information needed
35
-
36
- ## Training procedure
37
-
38
- ### Training hyperparameters
39
-
40
- The following hyperparameters were used during training:
41
  - learning_rate: 5e-07
42
  - train_batch_size: 8
43
  - eval_batch_size: 8
@@ -48,7 +47,7 @@ The following hyperparameters were used during training:
48
  - num_epochs: 1
49
  - mixed_precision_training: Native AMP
50
 
51
- ### Training results
52
 
53
  | Training Loss | Epoch | Step | Validation Loss | Wer |
54
  |:-------------:|:-----:|:----:|:---------------:|:-------:|
 
15
 
16
  # ꡬ음μž₯μ•  ν™˜μžλ₯Ό μœ„ν•œ μŒμ„±μΈμ‹ λͺ¨λΈ
17
 
18
+ ## ν”„λ‘œμ νŠΈ 정보
19
+ μž¬λ‹¨λ²•μΈ λ―Έλž˜μ™€ μ†Œν”„νŠΈμ›¨μ–΄μ™€ ν•¨κ»˜ν•˜λŠ” 제 3νšŒμ•„μ΄λ””μ–΄ 곡λͺ¨μ „
20
 
21
+ ## ν”„λ‘œμ νŠΈ λͺ…
22
+ "ꡬ음μž₯μ•  μŒμ„± 데이터λ₯Ό ν™œμš©ν•œ κ³ λ Ή ν™˜μžμ˜ μ˜μ‚¬μ†Œν†΅ κ°œμ„ λ°©μ•ˆ"
23
 
24
+ ## λͺ¨λΈ μ„€λͺ…
25
+ - **openai/whisper-large-v3**에 λŒ€ν•œ νŒŒμΈνŠœλ‹ λͺ¨λΈ
26
+ - λ³Έ λͺ¨λΈμ€ "ꡬ음μž₯μ•  μŒμ„± 데이터λ₯Ό ν™œμš©ν•œ κ³ λ Ή ν™˜μžμ˜ μ˜μ‚¬μ†Œν†΅ κ°œμ„ λ°©μ•ˆ" ν”„λ‘œμ νŠΈμ˜ ꡬ음μž₯μ• ν™˜μžλ“€μ— λŒ€ν•œ ν•œκ΅­μ–΄ μŒμ„±μΈμ‹ λͺ¨λΈμž„. OpenAI의 Whisper λͺ¨λΈμ„ νŒŒμΈνŠœλ‹ ν•˜μ—¬ ꡬ음μž₯μ• μ˜ μŒμ„±μ  νŠΉμ„±μ„ λ°˜μ˜ν•œ λͺ¨λΈμ„ κ΅¬μΆ•ν•˜μ˜€μŒ.
27
+ - 였λ₯Έμͺ½ "Inference API"λ₯Ό 톡해 μŒμ„±μΈμ‹ λͺ¨λΈμ„ ν…ŒμŠ€νŠΈ ν•΄λ³Ό 수 μžˆμŠ΅λ‹ˆλ‹€.
28
 
29
+ ## ν•™μŠ΅ λͺ¨λΈ
30
+ - **Paper**: Radford, A., Kim, J. W., Xu, T., Brockman, G., McLeavey, C., & Sutskever, I. (2023, July). Robust speech recognition via large-scale weak supervision. In International Conference on Machine Learning (pp. 28492-28518). PMLR.
31
+ - **URL**: https://proceedings.mlr.press/v202/radford23a.html
32
 
33
+ ## ν•™μŠ΅ 데이터
34
 
35
+ - AIHub "ꡬ음μž₯μ•  μŒμ„± 데이터" (KOR)
36
+ - URL: https://aihub.or.kr/aihubdata/data/view.do?currMenu=115&topMenu=100&aihubDataSe=data&dataSetSn=608
37
 
38
+ ### ν•™μŠ΅ νŒŒλΌλ―Έν„°
39
 
 
 
 
 
 
 
 
40
  - learning_rate: 5e-07
41
  - train_batch_size: 8
42
  - eval_batch_size: 8
 
47
  - num_epochs: 1
48
  - mixed_precision_training: Native AMP
49
 
50
+ ### ν•™μŠ΅ κ²°κ³Ό
51
 
52
  | Training Loss | Epoch | Step | Validation Loss | Wer |
53
  |:-------------:|:-----:|:----:|:---------------:|:-------:|