File size: 2,637 Bytes
cca1010
 
 
 
 
 
 
 
 
 
 
 
c3f648e
 
 
 
96e1801
cca1010
 
 
 
6c9f9f2
96e1801
 
 
 
 
 
 
 
 
 
 
c237cb0
c3f648e
 
96e1801
f576da6
cca1010
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
c2307eb
96e1801
cca1010
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
c3f648e
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
---
library_name: transformers
license: apache-2.0
base_model: openai/whisper-small
tags:
- generated_from_trainer
metrics:
- wer
- bleu
model-index:
- name: whisper-small-FLEURS-GL-EN
  results: []
datasets:
- juanjucm/FLEURS-SpeechT-GL-EN
language:
- gl
- en
---

# whisper-small-FLEURS-GL-EN

This model is a fine-tuned version of [openai/whisper-small](https://huggingface.co/openai/whisper-small) trained on [juanjucm/FLEURS-SpeechT-GL-EN](https://huggingface.co/datasets/juanjucm/FLEURS-SpeechT-GL-EN)
for **Galician-to-English Text to Speech Translation** task. It takes galician speech audios as input and generates the correspondant translated transcription in English.

The motivation behind this work is to increase the visibility of the Galician language, making it more accessible for non-Galician speakers to understand and engage with Galician audio content.

This model was developed during a 3-week Speech Translation workshop organised by [Yasmin Moslem](https://huggingface.co/ymoslem).


### Performance and training details

Baseline model achieved a BLEU score of **16.0** on the evaluation dataset.

After fine-tuning, it achieves the following results on the evaluation set:
- Loss: 1.6607
- Wer: 67.1683
- **BLEU: 22.6201**
- **ChrF++: 46.11**

The following hyperparameters were used during training:
- learning_rate: 1.25e-05
- train_batch_size: 16
- eval_batch_size: 8
- seed: 42
- distributed_type: multi-GPU
- num_devices: 2
- total_train_batch_size: 32
- total_eval_batch_size: 16
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 8
- mixed_precision_training: Native AMP

### Training results

We used [BLEU Score](https://en.wikipedia.org/wiki/BLEU) as our reference translation metric for selecting the best checkpoint after training.

| Training Loss | Epoch | Step | Validation Loss | Wer     | Bleu    |
|:-------------:|:-----:|:----:|:---------------:|:-------:|:-------:|
| 1.3189        | 1.0   | 86   | 1.6608          | 67.1683 | 22.6201 |
| 0.6613        | 2.0   | 172  | 1.6643          | 68.5990 | 21.1576 |
| 0.3492        | 3.0   | 258  | 1.7873          | 69.7046 | 20.7371 |
| 0.1416        | 4.0   | 344  | 1.9098          | 69.9090 | 20.5952 |
| 0.0974        | 5.0   | 430  | 2.0487          | 70.0948 | 20.6740 |
| 0.061         | 6.0   | 516  | 2.1565          | 73.4578 | 19.2411 |
| 0.0384        | 7.0   | 602  | 2.2107          | 73.6622 | 19.5413 |
| 0.0203        | 8.0   | 688  | 2.2476          | 73.9874 | 19.4512 |


### Framework versions

- Transformers 4.45.1
- Pytorch 2.4.1+cu121
- Datasets 3.0.1
- Tokenizers 0.20.0