File size: 5,405 Bytes
0edebf9
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
---
license: apache-2.0
tags:
- generated_from_trainer
metrics:
- wer
model-index:
- name: wav2vec2LugandaASR
  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. -->

# wav2vec2LugandaASR

This model is a fine-tuned version of [facebook/wav2vec2-large-xlsr-53](https://huggingface.co/facebook/wav2vec2-large-xlsr-53) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 0.6798
- Wer: 0.5291

## 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: 0.0003
- train_batch_size: 16
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 2
- 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: 500
- num_epochs: 130

### Training results

| Training Loss | Epoch  | Step | Validation Loss | Wer    |
|:-------------:|:------:|:----:|:---------------:|:------:|
| 2.8406        | 1.94   | 100  | 2.8577          | 0.9993 |
| 2.7812        | 3.88   | 200  | 2.8315          | 0.9993 |
| 1.1352        | 5.83   | 300  | 1.0099          | 1.1813 |
| 0.5333        | 7.77   | 400  | 0.5782          | 0.7937 |
| 0.3341        | 9.71   | 500  | 0.5899          | 0.7265 |
| 0.2432        | 11.65  | 600  | 0.5352          | 0.7162 |
| 0.2146        | 13.59  | 700  | 0.5439          | 0.6466 |
| 0.1998        | 15.53  | 800  | 0.5865          | 0.6618 |
| 0.1576        | 17.48  | 900  | 0.5598          | 0.6309 |
| 0.1665        | 19.42  | 1000 | 0.5400          | 0.6135 |
| 0.1191        | 21.36  | 1100 | 0.5496          | 0.6004 |
| 0.1038        | 23.3   | 1200 | 0.6248          | 0.6084 |
| 0.104         | 25.24  | 1300 | 0.5517          | 0.5934 |
| 0.1025        | 27.18  | 1400 | 0.5933          | 0.6008 |
| 0.1024        | 29.13  | 1500 | 0.5693          | 0.5901 |
| 0.0935        | 31.07  | 1600 | 0.5842          | 0.5899 |
| 0.0851        | 33.01  | 1700 | 0.6291          | 0.6086 |
| 0.0773        | 34.95  | 1800 | 0.6138          | 0.5812 |
| 0.0873        | 36.89  | 1900 | 0.5944          | 0.5729 |
| 0.0634        | 38.83  | 2000 | 0.6180          | 0.5807 |
| 0.0631        | 40.78  | 2100 | 0.5904          | 0.5704 |
| 0.0709        | 42.72  | 2200 | 0.5855          | 0.5791 |
| 0.0576        | 44.66  | 2300 | 0.6096          | 0.5789 |
| 0.0605        | 46.6   | 2400 | 0.5749          | 0.5617 |
| 0.0795        | 48.54  | 2500 | 0.5974          | 0.5749 |
| 0.0543        | 50.49  | 2600 | 0.6386          | 0.5754 |
| 0.0531        | 52.43  | 2700 | 0.6469          | 0.5794 |
| 0.0554        | 54.37  | 2800 | 0.6340          | 0.5555 |
| 0.0515        | 56.31  | 2900 | 0.6500          | 0.5762 |
| 0.0439        | 58.25  | 3000 | 0.6376          | 0.5758 |
| 0.0461        | 60.19  | 3100 | 0.6265          | 0.5711 |
| 0.0479        | 62.14  | 3200 | 0.6230          | 0.5707 |
| 0.039         | 64.08  | 3300 | 0.6337          | 0.5584 |
| 0.0397        | 66.02  | 3400 | 0.6347          | 0.5736 |
| 0.0509        | 67.96  | 3500 | 0.5946          | 0.5483 |
| 0.0471        | 69.9   | 3600 | 0.6355          | 0.5584 |
| 0.0481        | 71.84  | 3700 | 0.6514          | 0.5559 |
| 0.0484        | 73.79  | 3800 | 0.6373          | 0.5566 |
| 0.041         | 75.73  | 3900 | 0.6736          | 0.5646 |
| 0.0349        | 77.67  | 4000 | 0.6375          | 0.5622 |
| 0.0349        | 79.61  | 4100 | 0.6158          | 0.5506 |
| 0.0273        | 81.55  | 4200 | 0.6914          | 0.5666 |
| 0.029         | 83.5   | 4300 | 0.6361          | 0.5399 |
| 0.0353        | 85.44  | 4400 | 0.6397          | 0.5584 |
| 0.0289        | 87.38  | 4500 | 0.6554          | 0.5499 |
| 0.0257        | 89.32  | 4600 | 0.6676          | 0.5557 |
| 0.0403        | 91.26  | 4700 | 0.6440          | 0.5584 |
| 0.0361        | 93.2   | 4800 | 0.6587          | 0.5521 |
| 0.0304        | 95.15  | 4900 | 0.6837          | 0.5454 |
| 0.0289        | 97.09  | 5000 | 0.6684          | 0.5370 |
| 0.0282        | 99.03  | 5100 | 0.6556          | 0.5296 |
| 0.0302        | 100.97 | 5200 | 0.6833          | 0.5394 |
| 0.0196        | 102.91 | 5300 | 0.6837          | 0.5291 |
| 0.0255        | 104.85 | 5400 | 0.6644          | 0.5374 |
| 0.0209        | 106.8  | 5500 | 0.6700          | 0.5289 |
| 0.0243        | 108.74 | 5600 | 0.6835          | 0.5338 |
| 0.0203        | 110.68 | 5700 | 0.6850          | 0.5410 |
| 0.0237        | 112.62 | 5800 | 0.6561          | 0.5349 |
| 0.0251        | 114.56 | 5900 | 0.6776          | 0.5298 |
| 0.0177        | 116.5  | 6000 | 0.6748          | 0.5282 |
| 0.0232        | 118.45 | 6100 | 0.6767          | 0.5296 |
| 0.0257        | 120.39 | 6200 | 0.6793          | 0.5320 |
| 0.0194        | 122.33 | 6300 | 0.6804          | 0.5303 |
| 0.0304        | 124.27 | 6400 | 0.6798          | 0.5287 |
| 0.0251        | 126.21 | 6500 | 0.6798          | 0.5291 |
| 0.0201        | 128.16 | 6600 | 0.6798          | 0.5291 |


### Framework versions

- Transformers 4.30.2
- Pytorch 2.0.1+cu118
- Datasets 2.13.0
- Tokenizers 0.13.3