File size: 4,566 Bytes
578839e
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
---
language:
- ko
license: apache-2.0
base_model: openai/whisper-small
tags:
- hf-asr-leaderboard
- generated_from_trainer
datasets:
- arrow
metrics:
- wer
model-index:
- name: whisper-kor_noising_full
  results:
  - task:
      name: Automatic Speech Recognition
      type: automatic-speech-recognition
    dataset:
      name: whisper-kor_noising_full
      type: arrow
      config: default
      split: train
      args: 'config: ko, split: valid'
    metrics:
    - name: Wer
      type: wer
      value: 15.237651444547994
---

<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->

# whisper-kor_noising_full

This model is a fine-tuned version of [openai/whisper-small](https://huggingface.co/openai/whisper-small) on the whisper-kor_noising_full dataset.
It achieves the following results on the evaluation set:
- Loss: 0.1966
- Wer: 15.2377
- Cer: 7.1689

## Model description

More information needed

## Intended uses & limitations

More information needed

## Training and evaluation data

More information needed

## Training procedure

### Training hyperparameters

The following hyperparameters were used during training:
- learning_rate: 1e-05
- train_batch_size: 16
- eval_batch_size: 16
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 500
- training_steps: 4000

### Training results

| Training Loss | Epoch | Step | Validation Loss | Wer     | Cer    |
|:-------------:|:-----:|:----:|:---------------:|:-------:|:------:|
| 0.1965        | 0.05  | 100  | 0.1848          | 13.3271 | 5.5988 |
| 0.2196        | 0.09  | 200  | 0.1880          | 13.7310 | 5.6454 |
| 0.2601        | 0.14  | 300  | 0.1932          | 14.8493 | 6.1532 |
| 0.2118        | 0.18  | 400  | 0.2005          | 15.8434 | 6.5002 |
| 0.2784        | 0.23  | 500  | 0.2088          | 16.0298 | 6.7160 |
| 0.2421        | 0.28  | 600  | 0.2105          | 16.0920 | 6.7160 |
| 0.2209        | 0.32  | 700  | 0.2159          | 16.9773 | 7.0884 |
| 0.2426        | 0.37  | 800  | 0.2157          | 17.2258 | 7.1096 |
| 0.2429        | 0.42  | 900  | 0.2166          | 16.7754 | 6.9403 |
| 0.258         | 0.46  | 1000 | 0.2158          | 17.2569 | 7.0673 |
| 0.2605        | 0.51  | 1100 | 0.2135          | 16.5113 | 6.9784 |
| 0.2196        | 0.55  | 1200 | 0.2120          | 16.7443 | 6.8261 |
| 0.2423        | 0.6   | 1300 | 0.2163          | 16.8841 | 7.0884 |
| 0.2389        | 0.65  | 1400 | 0.2138          | 16.6201 | 7.0419 |
| 0.2314        | 0.69  | 1500 | 0.2149          | 16.8531 | 6.8599 |
| 0.2509        | 0.74  | 1600 | 0.2126          | 17.2103 | 7.8206 |
| 0.2329        | 0.78  | 1700 | 0.2103          | 16.0764 | 6.7457 |
| 0.2504        | 0.83  | 1800 | 0.2092          | 15.8590 | 6.6526 |
| 0.2632        | 0.88  | 1900 | 0.2107          | 16.2783 | 6.8726 |
| 0.2374        | 0.92  | 2000 | 0.2091          | 16.3249 | 6.7245 |
| 0.2625        | 0.97  | 2100 | 0.2057          | 15.7658 | 6.5425 |
| 0.1471        | 1.02  | 2200 | 0.2052          | 15.8434 | 6.5129 |
| 0.1541        | 1.06  | 2300 | 0.2069          | 16.3249 | 6.7457 |
| 0.1301        | 1.11  | 2400 | 0.2042          | 15.9211 | 6.4917 |
| 0.1674        | 1.15  | 2500 | 0.2058          | 15.3153 | 6.4240 |
| 0.1435        | 1.2   | 2600 | 0.2060          | 15.6726 | 6.5044 |
| 0.1352        | 1.25  | 2700 | 0.2040          | 15.2998 | 6.3902 |
| 0.1258        | 1.29  | 2800 | 0.2019          | 15.1600 | 6.2971 |
| 0.1273        | 1.34  | 2900 | 0.2025          | 15.6881 | 6.4875 |
| 0.1527        | 1.39  | 3000 | 0.2031          | 15.7036 | 6.5044 |
| 0.1371        | 1.43  | 3100 | 0.2011          | 15.3308 | 6.3309 |
| 0.1247        | 1.48  | 3200 | 0.2003          | 15.2842 | 6.3521 |
| 0.1376        | 1.52  | 3300 | 0.1987          | 15.4551 | 7.2366 |
| 0.1194        | 1.57  | 3400 | 0.1999          | 15.5949 | 7.2704 |
| 0.144         | 1.62  | 3500 | 0.1983          | 14.9425 | 6.2886 |
| 0.1387        | 1.66  | 3600 | 0.1979          | 14.9425 | 6.2082 |
| 0.1372        | 1.71  | 3700 | 0.1979          | 15.3464 | 7.1435 |
| 0.1513        | 1.75  | 3800 | 0.1972          | 15.1445 | 7.0334 |
| 0.134         | 1.8   | 3900 | 0.1970          | 15.2377 | 7.1646 |
| 0.1165        | 1.85  | 4000 | 0.1966          | 15.2377 | 7.1689 |


### Framework versions

- Transformers 4.33.2
- Pytorch 2.1.0+cu121
- Datasets 2.14.5
- Tokenizers 0.13.3