File size: 2,303 Bytes
64a67e2
2dd1d28
 
64a67e2
 
2dd1d28
64a67e2
 
2dd1d28
64a67e2
 
 
2dd1d28
64a67e2
 
 
 
 
2dd1d28
 
64a67e2
 
 
 
 
 
2dd1d28
64a67e2
 
 
 
 
2dd1d28
64a67e2
2dd1d28
64a67e2
2dd1d28
 
64a67e2
 
 
 
 
 
 
93b957f
64a67e2
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
6d7e844
 
 
 
 
64a67e2
 
 
 
 
 
 
 
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
---
language:
- eu
license: apache-2.0
tags:
- whisper-event
- generated_from_trainer
datasets:
- mozilla-foundation/common_voice_13_0
metrics:
- wer
model-index:
- name: Whisper Small Basque
  results:
  - task:
      name: Automatic Speech Recognition
      type: automatic-speech-recognition
    dataset:
      name: mozilla-foundation/common_voice_13_0 eu
      type: mozilla-foundation/common_voice_13_0
      config: eu
      split: test
      args: eu
    metrics:
    - name: Wer
      type: wer
      value: 18.775568066750374
---

<!-- 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. -->

# Whisper Small Basque

This model is a fine-tuned version of [openai/whisper-small](https://huggingface.co/openai/whisper-small) on the mozilla-foundation/common_voice_13_0 eu dataset.
It achieves the following results on the evaluation set:
- Loss: 0.3812
- Wer: 18.7756

## Model description

More information needed

## Intended uses & limitations

If you need to use this model with [whisper.cpp](https://github.com/ggerganov/whisper.cpp), you can download the ggml file: [ggml-small-eu.bin](https://huggingface.co/xezpeleta/whisper-small-eu/blob/main/ggml-small.eu.bin)

## Training and evaluation data

More information needed

## Training procedure

### Training hyperparameters

The following hyperparameters were used during training:
- learning_rate: 1e-05
- train_batch_size: 32
- eval_batch_size: 16
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 500
- training_steps: 5000
- mixed_precision_training: Native AMP

### Training results

| Training Loss | Epoch | Step | Validation Loss | Wer     |
|:-------------:|:-----:|:----:|:---------------:|:-------:|
| 0.1413        | 2.04  | 1000 | 0.3178          | 22.0139 |
| 0.0181        | 4.07  | 2000 | 0.3376          | 20.2864 |
| 0.0044        | 7.02  | 3000 | 0.3603          | 18.8768 |
| 0.0016        | 9.06  | 4000 | 0.3812          | 18.7756 |
| 0.0012        | 12.01 | 5000 | 0.3914          | 18.8302 |


### Framework versions

- Transformers 4.26.0.dev0
- Pytorch 1.13.1+cu117
- Datasets 2.8.1.dev0
- Tokenizers 0.13.2