File size: 5,447 Bytes
1886af1
 
44025aa
 
1886af1
 
 
44025aa
1886af1
44025aa
 
1886af1
 
 
44025aa
 
 
 
 
 
 
 
 
 
 
 
1886af1
 
 
 
 
44025aa
1886af1
44025aa
1886af1
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
---
library_name: transformers
language:
- ko
license: apache-2.0
base_model: openai/whisper-small
tags:
- whisper-event
- generated_from_trainer
datasets:
- GGarri/241113_newdata
metrics:
- wer
model-index:
- name: Whisper Small ko
  results:
  - task:
      name: Automatic Speech Recognition
      type: automatic-speech-recognition
    dataset:
      name: customdata
      type: GGarri/241113_newdata
    metrics:
    - name: Wer
      type: wer
      value: 0.8156606851549755
---

<!-- 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 Small ko

This model is a fine-tuned version of [openai/whisper-small](https://huggingface.co/openai/whisper-small) on the customdata dataset.
It achieves the following results on the evaluation set:
- Loss: 0.0498
- Cer: 1.1070
- Wer: 0.8157

## 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: 32
- eval_batch_size: 8
- seed: 42
- optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 500
- training_steps: 5000
- mixed_precision_training: Native AMP

### Training results

| Training Loss | Epoch   | Step | Validation Loss | Cer     | Wer     |
|:-------------:|:-------:|:----:|:---------------:|:-------:|:-------:|
| 1.1429        | 1.5625  | 100  | 0.8829          | 14.7984 | 14.5304 |
| 0.3401        | 3.125   | 200  | 0.2637          | 2.0625  | 1.7828  |
| 0.0413        | 4.6875  | 300  | 0.0599          | 1.5498  | 1.3167  |
| 0.0163        | 6.25    | 400  | 0.0462          | 1.2818  | 0.9904  |
| 0.0127        | 7.8125  | 500  | 0.0517          | 1.5265  | 1.1885  |
| 0.0065        | 9.375   | 600  | 0.0402          | 1.5031  | 1.0487  |
| 0.0028        | 10.9375 | 700  | 0.0396          | 1.7012  | 1.3167  |
| 0.001         | 12.5    | 800  | 0.0406          | 1.5148  | 1.1186  |
| 0.0004        | 14.0625 | 900  | 0.0405          | 1.4216  | 1.0371  |
| 0.0005        | 15.625  | 1000 | 0.0424          | 1.5847  | 1.1885  |
| 0.0001        | 17.1875 | 1100 | 0.0425          | 1.2701  | 0.9788  |
| 0.0001        | 18.75   | 1200 | 0.0429          | 1.3051  | 1.0137  |
| 0.0001        | 20.3125 | 1300 | 0.0432          | 1.2701  | 0.9788  |
| 0.0001        | 21.875  | 1400 | 0.0436          | 1.2818  | 0.9904  |
| 0.0001        | 23.4375 | 1500 | 0.0439          | 1.2934  | 1.0021  |
| 0.0001        | 25.0    | 1600 | 0.0441          | 1.2934  | 1.0021  |
| 0.0001        | 26.5625 | 1700 | 0.0443          | 1.2934  | 1.0021  |
| 0.0001        | 28.125  | 1800 | 0.0446          | 1.2934  | 1.0021  |
| 0.0001        | 29.6875 | 1900 | 0.0448          | 1.2818  | 0.9904  |
| 0.0001        | 31.25   | 2000 | 0.0449          | 1.2002  | 0.9089  |
| 0.0001        | 32.8125 | 2100 | 0.0454          | 1.2002  | 0.9089  |
| 0.0001        | 34.375  | 2200 | 0.0458          | 1.2002  | 0.9089  |
| 0.0           | 35.9375 | 2300 | 0.0461          | 1.2002  | 0.9089  |
| 0.0           | 37.5    | 2400 | 0.0463          | 1.1769  | 0.8856  |
| 0.0           | 39.0625 | 2500 | 0.0465          | 1.1769  | 0.8856  |
| 0.0           | 40.625  | 2600 | 0.0467          | 1.1536  | 0.8623  |
| 0.0           | 42.1875 | 2700 | 0.0469          | 1.1303  | 0.8390  |
| 0.0           | 43.75   | 2800 | 0.0471          | 1.1536  | 0.8623  |
| 0.0           | 45.3125 | 2900 | 0.0473          | 1.1536  | 0.8623  |
| 0.0           | 46.875  | 3000 | 0.0474          | 1.1536  | 0.8623  |
| 0.0           | 48.4375 | 3100 | 0.0476          | 1.1536  | 0.8623  |
| 0.0           | 50.0    | 3200 | 0.0477          | 1.1303  | 0.8390  |
| 0.0           | 51.5625 | 3300 | 0.0478          | 1.1419  | 0.8506  |
| 0.0           | 53.125  | 3400 | 0.0479          | 1.1186  | 0.8273  |
| 0.0           | 54.6875 | 3500 | 0.0481          | 1.1186  | 0.8273  |
| 0.0           | 56.25   | 3600 | 0.0482          | 1.1186  | 0.8273  |
| 0.0           | 57.8125 | 3700 | 0.0483          | 1.1186  | 0.8273  |
| 0.0           | 59.375  | 3800 | 0.0484          | 1.1070  | 0.8157  |
| 0.0           | 60.9375 | 3900 | 0.0485          | 1.1070  | 0.8157  |
| 0.0           | 62.5    | 4000 | 0.0487          | 1.1070  | 0.8157  |
| 0.0           | 64.0625 | 4100 | 0.0490          | 1.1070  | 0.8157  |
| 0.0           | 65.625  | 4200 | 0.0492          | 1.1070  | 0.8157  |
| 0.0           | 67.1875 | 4300 | 0.0494          | 1.1070  | 0.8157  |
| 0.0           | 68.75   | 4400 | 0.0495          | 1.1070  | 0.8157  |
| 0.0           | 70.3125 | 4500 | 0.0496          | 1.1070  | 0.8157  |
| 0.0           | 71.875  | 4600 | 0.0497          | 1.1070  | 0.8157  |
| 0.0           | 73.4375 | 4700 | 0.0497          | 1.1070  | 0.8157  |
| 0.0           | 75.0    | 4800 | 0.0497          | 1.1070  | 0.8157  |
| 0.0           | 76.5625 | 4900 | 0.0498          | 1.1070  | 0.8157  |
| 0.0           | 78.125  | 5000 | 0.0498          | 1.1070  | 0.8157  |


### Framework versions

- Transformers 4.46.2
- Pytorch 2.4.0
- Datasets 2.18.0
- Tokenizers 0.20.3