File size: 4,434 Bytes
d9bbcb5
 
 
0df87a6
 
d9bbcb5
 
 
 
 
 
 
 
 
 
 
 
 
0df87a6
d9bbcb5
0df87a6
 
d9bbcb5
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
---
license: apache-2.0
tags:
- automatic-speech-recognition
- hts98/original_ver1.2
- generated_from_trainer
metrics:
- wer
model-index:
- name: wav2vec2-ver2.0
  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-ver2.0

This model is a fine-tuned version of [facebook/wav2vec2-base](https://huggingface.co/facebook/wav2vec2-base) on the HTS98/ORIGINAL_VER1.2 - NA dataset.
It achieves the following results on the evaluation set:
- Loss: 0.9210
- Wer: 0.3901

## 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: 8
- eval_batch_size: 16
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 32
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 200
- num_epochs: 50.0

### Training results

| Training Loss | Epoch | Step | Validation Loss | Wer    |
|:-------------:|:-----:|:----:|:---------------:|:------:|
| No log        | 0.99  | 104  | 4.0281          | 1.0    |
| No log        | 2.0   | 209  | 3.4058          | 1.0    |
| No log        | 2.99  | 312  | 3.3457          | 1.0    |
| No log        | 4.0   | 417  | 3.3384          | 1.0    |
| 3.351         | 5.0   | 522  | 3.3195          | 1.0    |
| 3.351         | 6.0   | 627  | 3.2774          | 1.0000 |
| 3.351         | 6.99  | 731  | 2.7498          | 1.0184 |
| 3.351         | 8.0   | 836  | 2.0695          | 0.9268 |
| 3.351         | 9.0   | 941  | 1.6666          | 0.8103 |
| 2.5482        | 10.0  | 1046 | 1.4119          | 0.7014 |
| 2.5482        | 10.99 | 1150 | 1.2480          | 0.6408 |
| 2.5482        | 12.0  | 1255 | 1.1397          | 0.5814 |
| 2.5482        | 13.0  | 1360 | 1.0593          | 0.5381 |
| 2.5482        | 14.0  | 1465 | 1.0060          | 0.5099 |
| 1.1172        | 14.99 | 1569 | 0.9678          | 0.4830 |
| 1.1172        | 16.0  | 1674 | 0.9379          | 0.4692 |
| 1.1172        | 17.0  | 1779 | 0.9127          | 0.4618 |
| 1.1172        | 18.0  | 1884 | 0.8923          | 0.4352 |
| 1.1172        | 18.99 | 1988 | 0.8827          | 0.4254 |
| 0.7161        | 20.0  | 2093 | 0.8722          | 0.4304 |
| 0.7161        | 21.0  | 2198 | 0.8755          | 0.4142 |
| 0.7161        | 22.0  | 2303 | 0.8680          | 0.4157 |
| 0.7161        | 22.99 | 2407 | 0.8705          | 0.4116 |
| 0.5338        | 24.0  | 2512 | 0.8611          | 0.4039 |
| 0.5338        | 25.0  | 2617 | 0.8716          | 0.3995 |
| 0.5338        | 26.0  | 2722 | 0.8721          | 0.4037 |
| 0.5338        | 26.99 | 2826 | 0.8809          | 0.3973 |
| 0.5338        | 28.0  | 2931 | 0.9037          | 0.3938 |
| 0.4299        | 29.0  | 3036 | 0.9119          | 0.3903 |
| 0.4299        | 30.0  | 3141 | 0.9117          | 0.3912 |
| 0.4299        | 30.99 | 3245 | 0.9027          | 0.3930 |
| 0.4299        | 31.99 | 3328 | 0.9240          | 0.3898 |
| 0.4299        | 33.0  | 3433 | 0.9337          | 0.3872 |
| 0.3491        | 34.0  | 3538 | 0.9210          | 0.3901 |
| 0.3491        | 35.0  | 3643 | 0.9309          | 0.3905 |
| 0.3491        | 35.99 | 3747 | 0.9528          | 0.3906 |
| 0.3491        | 37.0  | 3852 | 0.9506          | 0.3880 |
| 0.3491        | 38.0  | 3957 | 0.9607          | 0.3853 |
| 0.3195        | 39.0  | 4062 | 0.9567          | 0.3906 |
| 0.3195        | 39.99 | 4166 | 0.9632          | 0.3893 |
| 0.3195        | 41.0  | 4271 | 0.9797          | 0.3839 |
| 0.3195        | 42.0  | 4376 | 0.9819          | 0.3854 |
| 0.3195        | 43.0  | 4481 | 0.9686          | 0.3870 |
| 0.2892        | 43.99 | 4585 | 0.9808          | 0.3895 |
| 0.2892        | 45.0  | 4690 | 0.9857          | 0.3892 |
| 0.2892        | 46.0  | 4795 | 0.9959          | 0.3831 |
| 0.2892        | 47.0  | 4900 | 0.9959          | 0.3870 |
| 0.2705        | 47.99 | 5004 | 1.0028          | 0.3860 |
| 0.2705        | 49.0  | 5109 | 1.0019          | 0.3869 |
| 0.2705        | 49.86 | 5200 | 1.0050          | 0.3857 |


### Framework versions

- Transformers 4.31.0.dev0
- Pytorch 2.0.1+cu118
- Datasets 2.7.0
- Tokenizers 0.13.3