Hugging Face's logo

stevehoang9
/
whisper-small-vi

TensorBoard
Safetensors
Vietnamese
Generated from Trainer
Model card Files Metrics Community
stevehoang9 commited on
Commit
54df2b4
•
1 Parent(s): a0c01ea

update model card README.md

Browse files
Files changed (1) hide show
  1. README.md +19 -39
README.md CHANGED
@@ -1,39 +1,24 @@
1
  ---
2
- base_model: openai/whisper-small
3
- datasets:
4
- - stevehoang9/vietnamese_ASR
5
  language:
6
  - vi
7
- library_name: peft
8
  license: apache-2.0
9
- metrics:
10
- - wer
11
  tags:
12
  - generated_from_trainer
 
 
13
  model-index:
14
- - name: Whisper Small Vietnam - Steve Hoang
15
- results:
16
- - task:
17
- type: automatic-speech-recognition
18
- name: Automatic Speech Recognition
19
- dataset:
20
- name: VIVOS
21
- type: stevehoang9/vietnamese_ASR
22
- metrics:
23
- - type: wer
24
- value: 51.68350168350169
25
- name: Wer
26
  ---
27
 
28
  <!-- This model card has been generated automatically according to the information the Trainer had access to. You
29
  should probably proofread and complete it, then remove this comment. -->
30
 
31
- # Whisper Small Vietnam - Steve Hoang
32
 
33
- This model is a fine-tuned version of [openai/whisper-small](https://huggingface.co/openai/whisper-small) on the VIVOS dataset.
34
  It achieves the following results on the evaluation set:
35
- - Loss: 0.6333
36
- - Wer: 51.6835
37
 
38
  ## Model description
39
 
@@ -54,30 +39,25 @@ More information needed
54
  The following hyperparameters were used during training:
55
  - learning_rate: 1e-05
56
  - train_batch_size: 16
57
- - eval_batch_size: 8
58
  - seed: 42
59
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
60
- - lr_scheduler_type: linear
61
- - lr_scheduler_warmup_steps: 500
62
- - training_steps: 5000
63
- - mixed_precision_training: Native AMP
64
 
65
  ### Training results
66
 
67
- | Training Loss | Epoch | Step | Validation Loss | Wer |
68
- |:-------------:|:------:|:----:|:---------------:|:-------:|
69
- | 0.931 | 1.3717 | 1000 | 0.9599 | 74.7216 |
70
- | 0.7286 | 2.7435 | 2000 | 0.7709 | 61.5255 |
71
- | 0.6341 | 4.1152 | 3000 | 0.6854 | 55.5685 |
72
- | 0.6044 | 5.4870 | 4000 | 0.6458 | 52.8490 |
73
- | 0.6031 | 6.8587 | 5000 | 0.6333 | 51.6835 |
74
 
75
 
76
  ### Framework versions
77
 
78
- - PEFT 0.11.2.dev0
79
- - PEFT 0.11.1
80
- - Transformers 4.41.2
81
  - Pytorch 2.1.2
82
- - Datasets 2.19.2
83
- - Tokenizers 0.19.1
 
1
  ---
 
 
 
2
  language:
3
  - vi
 
4
  license: apache-2.0
 
 
5
  tags:
6
  - generated_from_trainer
7
+ datasets:
8
+ - JRHuy/vivos-fleurs
9
  model-index:
10
+ - name: Whisper Small Vietnamese - Steve Hoang
11
+ results: []
 
 
 
 
 
 
 
 
 
 
12
  ---
13
 
14
  <!-- This model card has been generated automatically according to the information the Trainer had access to. You
15
  should probably proofread and complete it, then remove this comment. -->
16
 
17
+ # Whisper Small Vietnamese - Steve Hoang
18
 
19
+ This model is a fine-tuned version of [openai/whisper-small](https://huggingface.co/openai/whisper-small) on the Vietnamese ASR dataset.
20
  It achieves the following results on the evaluation set:
21
+ - Loss: 1.5475
 
22
 
23
  ## Model description
24
 
 
39
  The following hyperparameters were used during training:
40
  - learning_rate: 1e-05
41
  - train_batch_size: 16
42
+ - eval_batch_size: 16
43
  - seed: 42
44
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
45
+ - lr_scheduler_type: constant_with_warmup
46
+ - lr_scheduler_warmup_steps: 50
47
+ - training_steps: 300
 
48
 
49
  ### Training results
50
 
51
+ | Training Loss | Epoch | Step | Validation Loss |
52
+ |:-------------:|:-----:|:----:|:---------------:|
53
+ | 2.5095 | 0.11 | 100 | 2.3242 |
54
+ | 2.0473 | 0.21 | 200 | 1.9481 |
55
+ | 1.6277 | 0.32 | 300 | 1.5475 |
 
 
56
 
57
 
58
  ### Framework versions
59
 
60
+ - Transformers 4.30.0
 
 
61
  - Pytorch 2.1.2
62
+ - Datasets 2.20.0
63
+ - Tokenizers 0.13.3