Hugging Face's logo

anhphuong
/
whisper_largev2_jp

Automatic Speech Recognition
Transformers
TensorBoard
Safetensors
Vietnamese
whisper
Generated from Trainer
Eval Results
Inference Endpoints
Model card Files Metrics Community
anhphuong commited on
Commit
7becd6d
•
1 Parent(s): d2e6b9e

End of training

Browse files
Files changed (6) hide show
  1. README.md +19 -19
  2. generation_config.json +76 -8
  3. model-00001-of-00002.safetensors +3 -0
  4. model-00002-of-00002.safetensors +3 -0
  5. model.safetensors.index.json +0 -0
  6. runs/Jul07_10-31-07_ad68762e8c0d/events.out.tfevents.1720348292.ad68762e8c0d.366.0 +2 -2
README.md CHANGED
@@ -1,40 +1,40 @@
1
  ---
2
  language:
3
  - vi
4
- base_model: openai/whisper-medium-ja-v2
5
  tags:
6
  - generated_from_trainer
7
  datasets:
8
- - mozilla-foundation/common_voice_11_0
9
  metrics:
10
  - wer
11
  model-index:
12
- - name: Whisper Medium Vi - Anh Phuong
13
  results:
14
  - task:
15
  name: Automatic Speech Recognition
16
  type: automatic-speech-recognition
17
  dataset:
18
- name: Common Voice 11.0
19
- type: mozilla-foundation/common_voice_11_0
20
- config: ja
21
  split: None
22
  args: 'config: hi, split: test'
23
  metrics:
24
  - name: Wer
25
  type: wer
26
- value: 62.82245827010622
27
  ---
28
 
29
  <!-- This model card has been generated automatically according to the information the Trainer had access to. You
30
  should probably proofread and complete it, then remove this comment. -->
31
 
32
- # Whisper Medium Vi - Anh Phuong
33
 
34
- This model is a fine-tuned version of [openai/whisper-medium-ja-v2](https://huggingface.co/openai/whisper-medium-ja-v2) on the Common Voice 11.0 dataset.
35
  It achieves the following results on the evaluation set:
36
- - Loss: 0.3098
37
- - Wer: 62.8225
38
 
39
  ## Model description
40
 
@@ -65,14 +65,14 @@ The following hyperparameters were used during training:
65
 
66
  ### Training results
67
 
68
- | Training Loss | Epoch | Step | Validation Loss | Wer |
69
- |:-------------:|:------:|:----:|:---------------:|:-------:|
70
- | 0.1131 | 1.4556 | 1000 | 0.2257 | 68.5454 |
71
- | 0.0579 | 2.9112 | 2000 | 0.2363 | 65.5105 |
72
- | 0.0087 | 4.3668 | 3000 | 0.2685 | 65.1203 |
73
- | 0.003 | 5.8224 | 4000 | 0.2924 | 63.9931 |
74
- | 0.0007 | 7.2780 | 5000 | 0.3041 | 63.1043 |
75
- | 0.0005 | 8.7336 | 6000 | 0.3098 | 62.8225 |
76
 
77
 
78
  ### Framework versions
 
1
  ---
2
  language:
3
  - vi
4
+ base_model: openai/whisper-largev2-ja-v2
5
  tags:
6
  - generated_from_trainer
7
  datasets:
8
+ - google/fleurs
9
  metrics:
10
  - wer
11
  model-index:
12
+ - name: Whisper Large V2 Ja - Anh Phuong
13
  results:
14
  - task:
15
  name: Automatic Speech Recognition
16
  type: automatic-speech-recognition
17
  dataset:
18
+ name: Google fleurs
19
+ type: google/fleurs
20
+ config: ja_jp
21
  split: None
22
  args: 'config: hi, split: test'
23
  metrics:
24
  - name: Wer
25
  type: wer
26
+ value: 58.247168882323976
27
  ---
28
 
29
  <!-- This model card has been generated automatically according to the information the Trainer had access to. You
30
  should probably proofread and complete it, then remove this comment. -->
31
 
32
+ # Whisper Large V2 Ja - Anh Phuong
33
 
34
+ This model is a fine-tuned version of [openai/whisper-largev2-ja-v2](https://huggingface.co/openai/whisper-largev2-ja-v2) on the Google fleurs dataset.
35
  It achieves the following results on the evaluation set:
36
+ - Loss: 0.2626
37
+ - Wer: 58.2472
38
 
39
  ## Model description
40
 
 
65
 
66
  ### Training results
67
 
68
+ | Training Loss | Epoch | Step | Validation Loss | Wer |
69
+ |:-------------:|:-----:|:----:|:---------------:|:-------:|
70
+ | 0.004 | 6.25 | 1000 | 0.2030 | 61.0044 |
71
+ | 0.0022 | 12.5 | 2000 | 0.2081 | 60.6105 |
72
+ | 0.0002 | 18.75 | 3000 | 0.2401 | 58.7888 |
73
+ | 0.0001 | 25.0 | 4000 | 0.2531 | 58.6411 |
74
+ | 0.0001 | 31.25 | 5000 | 0.2598 | 58.2472 |
75
+ | 0.0001 | 37.5 | 6000 | 0.2626 | 58.2472 |
76
 
77
 
78
  ### Framework versions
generation_config.json CHANGED
@@ -1,28 +1,96 @@
1
  {
2
  "alignment_heads": [
 
 
 
 
3
  [
4
  13,
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
5
  15
6
  ],
7
  [
8
- 15,
 
 
 
 
9
  4
10
  ],
11
  [
12
- 15,
13
- 15
14
  ],
15
  [
16
- 16,
17
- 1
18
  ],
19
  [
20
- 20,
21
- 0
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
22
  ],
23
  [
24
  23,
25
- 4
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
26
  ]
27
  ],
28
  "begin_suppress_tokens": [
 
1
  {
2
  "alignment_heads": [
3
+ [
4
+ 10,
5
+ 12
6
+ ],
7
  [
8
  13,
9
+ 17
10
+ ],
11
+ [
12
+ 16,
13
+ 11
14
+ ],
15
+ [
16
+ 16,
17
+ 12
18
+ ],
19
+ [
20
+ 16,
21
+ 13
22
+ ],
23
+ [
24
+ 17,
25
  15
26
  ],
27
  [
28
+ 17,
29
+ 16
30
+ ],
31
+ [
32
+ 18,
33
  4
34
  ],
35
  [
36
+ 18,
37
+ 11
38
  ],
39
  [
40
+ 18,
41
+ 19
42
  ],
43
  [
44
+ 19,
45
+ 11
46
+ ],
47
+ [
48
+ 21,
49
+ 2
50
+ ],
51
+ [
52
+ 21,
53
+ 3
54
+ ],
55
+ [
56
+ 22,
57
+ 3
58
+ ],
59
+ [
60
+ 22,
61
+ 9
62
+ ],
63
+ [
64
+ 22,
65
+ 12
66
  ],
67
  [
68
  23,
69
+ 5
70
+ ],
71
+ [
72
+ 23,
73
+ 7
74
+ ],
75
+ [
76
+ 23,
77
+ 13
78
+ ],
79
+ [
80
+ 25,
81
+ 5
82
+ ],
83
+ [
84
+ 26,
85
+ 1
86
+ ],
87
+ [
88
+ 26,
89
+ 12
90
+ ],
91
+ [
92
+ 27,
93
+ 15
94
  ]
95
  ],
96
  "begin_suppress_tokens": [
model-00001-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:661cc2c0a823b4da986b198bfb75855a13845c5667e33682f45d41bd0a2fbf37
3
+ size 4992706480
model-00002-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5807485c0e54843cb73ac4fabe585a994f703594295db297a89fd361d400a5a6
3
+ size 1180663192
model.safetensors.index.json ADDED
The diff for this file is too large to render. See raw diff
 
runs/Jul07_10-31-07_ad68762e8c0d/events.out.tfevents.1720348292.ad68762e8c0d.366.0 CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:a2e28124823854ba11968e549ae0b7042bf1560d5d9c0d3c448dd6e454dce897
3
- size 59163
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:dcdf2fd3e3da40b3d2dc68340c4bb36f27a41147091691f6cabfd981fb5d2f0d
3
+ size 59517