File size: 3,137 Bytes
145b24a
 
 
 
ed087b1
 
145b24a
 
 
 
 
 
 
 
 
 
 
 
 
ed087b1
145b24a
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
---
license: apache-2.0
base_model: facebook/wav2vec2-xls-r-2b
tags:
- automatic-speech-recognition
- DewiBrynJones/banc-trawsgrifiadau-bangor-normalized
- generated_from_trainer
metrics:
- wer
model-index:
- name: wav2vec2-xls-r-2b-ft-btb-cy
  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-xls-r-2b-ft-btb-cy

This model is a fine-tuned version of [facebook/wav2vec2-xls-r-2b](https://huggingface.co/facebook/wav2vec2-xls-r-2b) on the DEWIBRYNJONES/BANC-TRAWSGRIFIADAU-BANGOR-NORMALIZED - DEFAULT dataset.
It achieves the following results on the evaluation set:
- Loss: 0.3903
- Wer: 0.2957

## 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
- training_steps: 2500
- mixed_precision_training: Native AMP

### Training results

| Training Loss | Epoch  | Step | Validation Loss | Wer    |
|:-------------:|:------:|:----:|:---------------:|:------:|
| No log        | 0.1414 | 100  | 1.2105          | 0.8709 |
| No log        | 0.2829 | 200  | 0.9787          | 0.6986 |
| No log        | 0.4243 | 300  | 1.1907          | 0.7127 |
| No log        | 0.5658 | 400  | 1.0559          | 0.7169 |
| 1.4456        | 0.7072 | 500  | 1.2106          | 0.7944 |
| 1.4456        | 0.8487 | 600  | 1.0232          | 0.7033 |
| 1.4456        | 0.9901 | 700  | 1.0387          | 0.7336 |
| 1.4456        | 1.1315 | 800  | 0.7234          | 0.5223 |
| 1.4456        | 1.2730 | 900  | 0.7242          | 0.5566 |
| 0.9155        | 1.4144 | 1000 | 0.7097          | 0.5259 |
| 0.9155        | 1.5559 | 1100 | 0.6368          | 0.4797 |
| 0.9155        | 1.6973 | 1200 | 0.6065          | 0.4653 |
| 0.9155        | 1.8388 | 1300 | 0.6207          | 0.4717 |
| 0.9155        | 1.9802 | 1400 | 0.5925          | 0.4707 |
| 0.7436        | 2.1216 | 1500 | 0.5382          | 0.4046 |
| 0.7436        | 2.2631 | 1600 | 0.5201          | 0.3996 |
| 0.7436        | 2.4045 | 1700 | 0.4883          | 0.3698 |
| 0.7436        | 2.5460 | 1800 | 0.4704          | 0.3659 |
| 0.7436        | 2.6874 | 1900 | 0.4443          | 0.3521 |
| 0.5645        | 2.8289 | 2000 | 0.4470          | 0.3476 |
| 0.5645        | 2.9703 | 2100 | 0.4192          | 0.3242 |
| 0.5645        | 3.1117 | 2200 | 0.4178          | 0.3161 |
| 0.5645        | 3.2532 | 2300 | 0.4122          | 0.3054 |
| 0.5645        | 3.3946 | 2400 | 0.3960          | 0.2990 |
| 0.4232        | 3.5361 | 2500 | 0.3903          | 0.2957 |


### Framework versions

- Transformers 4.40.2
- Pytorch 2.3.0+cu121
- Datasets 2.19.1
- Tokenizers 0.19.1