File size: 2,533 Bytes
1806b43
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
d7d406e
 
 
1806b43
 
 
 
 
 
 
 
 
 
 
 
 
d7d406e
 
 
 
 
1806b43
 
d7d406e
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
e27ff29
1806b43
 
 
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
---
language:
- zh
license: apache-2.0
tags:
- whisper-event
- generated_from_trainer
datasets:
- mozilla-foundation/common_voice_11_0
model-index:
- name: Whisper Small zh-HK - Alvin
  results:
  - task:
      name: Automatic Speech Recognition
      type: automatic-speech-recognition
    dataset:
      name: mozilla-foundation/common_voice_11_0 zh-HK
      type: mozilla-foundation/common_voice_11_0
      config: zh-HK
      split: test
      args: zh-HK
    metrics:
    - name: Cer
      type: cer
      value: 11.760
---
<!-- 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 zh-HK - Alvin

This model is a fine-tuned version of [openai/whisper-small](https://huggingface.co/openai/whisper-small) on the Common Voice 11.0 dataset.

## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
For training, three datasets were used:
- Common Voice 11 Canto Train Set
- CantoMap: Winterstein, Grégoire, Tang, Carmen and Lai, Regine (2020) "CantoMap: a Hong Kong Cantonese MapTask Corpus", in Proceedings of The 12th Language Resources and Evaluation Conference, Marseille: European Language Resources Association, p. 2899-2906.
- Cantonse-ASR: Yu, Tiezheng, Frieske, Rita, Xu, Peng, Cahyawijaya, Samuel, Yiu, Cheuk Tung, Lovenia, Holy, Dai, Wenliang, Barezi, Elham, Chen, Qifeng, Ma, Xiaojuan, Shi, Bertram, Fung, Pascale (2022) "Automatic Speech Recognition Datasets in Cantonese: A Survey and New Dataset", 2022. Link: https://arxiv.org/pdf/2201.02419.pdf

## Training procedure

## Training Hyperparameters
- learning_rate: 1e-5
- train_batch_size: 16 (on 2 GPUs)
- eval_batch_size: 8
- gradient_accumulation_steps: 2
- total_train_batch_size: 16x2x2=64
- 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 | Cer    |
|:-------------:|:-----:|:----:|:---------------:|:------:|
| 0.1106        | 0.66  | 1000 | 0.3294          | 14.638 |
| 0.0546        | 1.33  | 2000 | 0.2887          | 12.119 |
| 0.0293        | 2.01  | 3000 | 0.2727          | 11.646 |
| 0.0214        | 2.66  | 4000 | 0.2741          | 11.760 |
| 0.0919        | 3.32  | 5000 | 0.2747          | 11.463 |


### Framework versions