File size: 4,817 Bytes
a0c5fcc
 
 
 
 
 
 
 
 
 
 
 
 
6ad6e3b
a0c5fcc
6ad6e3b
 
a0c5fcc
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
e774553
 
a0c5fcc
 
 
 
e774553
6ad6e3b
a0c5fcc
 
 
6ad6e3b
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
a0c5fcc
 
 
 
 
 
 
 
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
---
tags:
- generated_from_trainer
model-index:
- name: wav2vec2-large-xls-r-300m-korean-d
  results: []
---

<!-- 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. -->

# wav2vec2-large-xls-r-300m-korean-d

This model was trained from scratch on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 5.0644
- Cer: 0.8255

## 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: 5e-05
- train_batch_size: 4
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 100
- num_epochs: 40

### Training results

| Training Loss | Epoch | Step | Validation Loss | Cer    |
|:-------------:|:-----:|:----:|:---------------:|:------:|
| 4.1174        | 0.66  | 50   | 5.1872          | 0.9986 |
| 4.0452        | 1.32  | 100  | 5.1870          | 0.9986 |
| 4.0499        | 1.97  | 150  | 5.2289          | 0.9986 |
| 4.0371        | 2.63  | 200  | 5.1608          | 0.9986 |
| 3.9664        | 3.29  | 250  | 5.1345          | 0.9977 |
| 3.991         | 3.95  | 300  | 5.1517          | 0.9968 |
| 3.9413        | 4.61  | 350  | 5.0673          | 0.9927 |
| 3.9433        | 5.26  | 400  | 5.0650          | 0.9823 |
| 3.8934        | 5.92  | 450  | 5.0518          | 0.9800 |
| 3.8646        | 6.58  | 500  | 5.0400          | 0.9823 |
| 3.8491        | 7.24  | 550  | 5.1012          | 0.9764 |
| 3.8725        | 7.89  | 600  | 5.0649          | 0.9855 |
| 3.7272        | 8.55  | 650  | 5.1139          | 0.9791 |
| 3.8121        | 9.21  | 700  | 5.0366          | 0.9409 |
| 3.7743        | 9.87  | 750  | 5.0990          | 0.9673 |
| 3.7207        | 10.53 | 800  | 5.0603          | 0.9278 |
| 3.7116        | 11.18 | 850  | 5.0920          | 0.9119 |
| 3.7163        | 11.84 | 900  | 5.0840          | 0.8996 |
| 3.657         | 12.5  | 950  | 5.0855          | 0.8928 |
| 3.6476        | 13.16 | 1000 | 5.0409          | 0.8851 |
| 3.645         | 13.82 | 1050 | 5.0704          | 0.9028 |
| 3.5882        | 14.47 | 1100 | 5.0391          | 0.8610 |
| 3.5773        | 15.13 | 1150 | 5.0805          | 0.8628 |
| 3.5681        | 15.79 | 1200 | 5.1300          | 0.8769 |
| 3.5611        | 16.45 | 1250 | 5.0740          | 0.8760 |
| 3.5221        | 17.11 | 1300 | 5.0698          | 0.8669 |
| 3.493         | 17.76 | 1350 | 5.0618          | 0.8455 |
| 3.5117        | 18.42 | 1400 | 5.0372          | 0.8433 |
| 3.4777        | 19.08 | 1450 | 5.0964          | 0.8642 |
| 3.4632        | 19.74 | 1500 | 5.0928          | 0.8623 |
| 3.4496        | 20.39 | 1550 | 5.1118          | 0.8710 |
| 3.4674        | 21.05 | 1600 | 5.0703          | 0.8392 |
| 3.431         | 21.71 | 1650 | 5.0514          | 0.8373 |
| 3.4115        | 22.37 | 1700 | 5.0611          | 0.8355 |
| 3.3808        | 23.03 | 1750 | 5.1055          | 0.8537 |
| 3.4101        | 23.68 | 1800 | 5.0532          | 0.8296 |
| 3.3852        | 24.34 | 1850 | 5.0646          | 0.8310 |
| 3.3533        | 25.0  | 1900 | 5.0684          | 0.8387 |
| 3.3591        | 25.66 | 1950 | 5.0581          | 0.8364 |
| 3.3437        | 26.32 | 2000 | 5.0565          | 0.8314 |
| 3.369         | 26.97 | 2050 | 5.0577          | 0.8364 |
| 3.3606        | 27.63 | 2100 | 5.0515          | 0.8237 |
| 3.3163        | 28.29 | 2150 | 5.0533          | 0.8278 |
| 3.3149        | 28.95 | 2200 | 5.0682          | 0.8292 |
| 3.3535        | 29.61 | 2250 | 5.0554          | 0.8274 |
| 3.2695        | 30.26 | 2300 | 5.0610          | 0.8242 |
| 3.2947        | 30.92 | 2350 | 5.0658          | 0.8255 |
| 3.3323        | 31.58 | 2400 | 5.0644          | 0.8255 |
| 3.2913        | 32.24 | 2450 | 5.0644          | 0.8255 |
| 3.3169        | 32.89 | 2500 | 5.0644          | 0.8255 |
| 3.3147        | 33.55 | 2550 | 5.0644          | 0.8255 |
| 3.3059        | 34.21 | 2600 | 5.0644          | 0.8255 |
| 3.3311        | 34.87 | 2650 | 5.0644          | 0.8255 |
| 3.286         | 35.53 | 2700 | 5.0644          | 0.8255 |
| 3.3842        | 36.18 | 2750 | 5.0644          | 0.8255 |
| 3.303         | 36.84 | 2800 | 5.0644          | 0.8255 |
| 3.2833        | 37.5  | 2850 | 5.0644          | 0.8255 |
| 3.3036        | 38.16 | 2900 | 5.0644          | 0.8255 |
| 3.3149        | 38.82 | 2950 | 5.0644          | 0.8255 |
| 3.2784        | 39.47 | 3000 | 5.0644          | 0.8255 |


### Framework versions

- Transformers 4.30.2
- Pytorch 2.0.1+cu118
- Datasets 2.13.1
- Tokenizers 0.13.3