Hugging Face's logo

ducha07
/
ASR-test

Automatic Speech Recognition
Transformers
Safetensors
Vietnamese
wav2vec2
Generated from Trainer
Eval Results (legacy)
Model card Files
xet
Community
ducha07 commited on
Commit
d10921f
·
1 Parent(s): bfd2cd3

End of training

Browse files
Files changed (3) hide show
  1. README.md +50 -8
  2. adapter.vie.safetensors +1 -1
  3. model.safetensors +1 -1
README.md CHANGED
@@ -10,7 +10,7 @@ datasets:
10
  metrics:
11
  - wer
12
  model-index:
13
- - name: ASR-test
14
  results:
15
  - task:
16
  name: Automatic Speech Recognition
@@ -21,18 +21,18 @@ model-index:
21
  metrics:
22
  - name: Wer
23
  type: wer
24
- value: 0.4502295081967213
25
  ---
26
 
27
  <!-- This model card has been generated automatically according to the information the Trainer had access to. You
28
  should probably proofread and complete it, then remove this comment. -->
29
 
30
- # ASR-test
31
 
32
  This model is a fine-tuned version of [facebook/mms-1b-all](https://huggingface.co/facebook/mms-1b-all) on the HTV news dataset.
33
  It achieves the following results on the evaluation set:
34
- - Loss: 0.8796
35
- - Wer: 0.4502
36
 
37
  ## Model description
38
 
@@ -58,19 +58,61 @@ The following hyperparameters were used during training:
58
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
59
  - lr_scheduler_type: linear
60
  - lr_scheduler_warmup_steps: 100
61
- - num_epochs: 1
62
  - mixed_precision_training: Native AMP
63
 
64
  ### Training results
65
 
66
  | Training Loss | Epoch | Step | Validation Loss | Wer |
67
  |:-------------:|:-----:|:----:|:---------------:|:------:|
68
- | 4.8401 | 0.92 | 100 | 0.8796 | 0.4502 |
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
69
 
70
 
71
  ### Framework versions
72
 
73
  - Transformers 4.37.0.dev0
74
- - Pytorch 2.1.0+cu118
75
  - Datasets 2.15.0
76
  - Tokenizers 0.15.0
 
10
  metrics:
11
  - wer
12
  model-index:
13
+ - name: ASR-test-1
14
  results:
15
  - task:
16
  name: Automatic Speech Recognition
 
21
  metrics:
22
  - name: Wer
23
  type: wer
24
+ value: 0.2882930019620667
25
  ---
26
 
27
  <!-- This model card has been generated automatically according to the information the Trainer had access to. You
28
  should probably proofread and complete it, then remove this comment. -->
29
 
30
+ # ASR-test-1
31
 
32
  This model is a fine-tuned version of [facebook/mms-1b-all](https://huggingface.co/facebook/mms-1b-all) on the HTV news dataset.
33
  It achieves the following results on the evaluation set:
34
+ - Loss: 0.5663
35
+ - Wer: 0.2883
36
 
37
  ## Model description
38
 
 
58
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
59
  - lr_scheduler_type: linear
60
  - lr_scheduler_warmup_steps: 100
61
+ - num_epochs: 40
62
  - mixed_precision_training: Native AMP
63
 
64
  ### Training results
65
 
66
  | Training Loss | Epoch | Step | Validation Loss | Wer |
67
  |:-------------:|:-----:|:----:|:---------------:|:------:|
68
+ | 4.772 | 0.92 | 100 | 0.8456 | 0.4411 |
69
+ | 1.1042 | 1.83 | 200 | 0.7041 | 0.4076 |
70
+ | 0.9814 | 2.75 | 300 | 0.7243 | 0.3782 |
71
+ | 0.9096 | 3.67 | 400 | 0.6771 | 0.3655 |
72
+ | 0.8823 | 4.59 | 500 | 0.6265 | 0.3627 |
73
+ | 0.8435 | 5.5 | 600 | 0.6200 | 0.3543 |
74
+ | 0.8157 | 6.42 | 700 | 0.6414 | 0.3417 |
75
+ | 0.822 | 7.34 | 800 | 0.5872 | 0.3431 |
76
+ | 0.7852 | 8.26 | 900 | 0.6012 | 0.3387 |
77
+ | 0.7533 | 9.17 | 1000 | 0.6023 | 0.3256 |
78
+ | 0.7609 | 10.09 | 1100 | 0.5837 | 0.3444 |
79
+ | 0.7568 | 11.01 | 1200 | 0.5791 | 0.3311 |
80
+ | 0.7091 | 11.93 | 1300 | 0.6227 | 0.3206 |
81
+ | 0.7098 | 12.84 | 1400 | 0.5766 | 0.3266 |
82
+ | 0.7006 | 13.76 | 1500 | 0.6084 | 0.3117 |
83
+ | 0.6673 | 14.68 | 1600 | 0.5857 | 0.3120 |
84
+ | 0.6832 | 15.6 | 1700 | 0.5754 | 0.3338 |
85
+ | 0.6646 | 16.51 | 1800 | 0.5963 | 0.3117 |
86
+ | 0.6524 | 17.43 | 1900 | 0.5816 | 0.3137 |
87
+ | 0.6385 | 18.35 | 2000 | 0.5691 | 0.3257 |
88
+ | 0.6433 | 19.27 | 2100 | 0.5929 | 0.3105 |
89
+ | 0.6129 | 20.18 | 2200 | 0.5709 | 0.3067 |
90
+ | 0.624 | 21.1 | 2300 | 0.5686 | 0.3168 |
91
+ | 0.6128 | 22.02 | 2400 | 0.5867 | 0.3080 |
92
+ | 0.584 | 22.94 | 2500 | 0.5680 | 0.3101 |
93
+ | 0.5956 | 23.85 | 2600 | 0.5611 | 0.3023 |
94
+ | 0.5825 | 24.77 | 2700 | 0.5821 | 0.2999 |
95
+ | 0.56 | 25.69 | 2800 | 0.5622 | 0.3012 |
96
+ | 0.56 | 26.61 | 2900 | 0.5590 | 0.3053 |
97
+ | 0.5523 | 27.52 | 3000 | 0.5758 | 0.2967 |
98
+ | 0.5335 | 28.44 | 3100 | 0.5649 | 0.3090 |
99
+ | 0.5686 | 29.36 | 3200 | 0.5703 | 0.2931 |
100
+ | 0.5488 | 30.28 | 3300 | 0.5709 | 0.2921 |
101
+ | 0.5249 | 31.19 | 3400 | 0.5646 | 0.2973 |
102
+ | 0.5278 | 32.11 | 3500 | 0.5628 | 0.2933 |
103
+ | 0.5252 | 33.03 | 3600 | 0.5663 | 0.2927 |
104
+ | 0.5092 | 33.94 | 3700 | 0.5618 | 0.2922 |
105
+ | 0.5099 | 34.86 | 3800 | 0.5616 | 0.2954 |
106
+ | 0.5031 | 35.78 | 3900 | 0.5670 | 0.2913 |
107
+ | 0.4959 | 36.7 | 4000 | 0.5679 | 0.2923 |
108
+ | 0.4936 | 37.61 | 4100 | 0.5675 | 0.2912 |
109
+ | 0.5012 | 38.53 | 4200 | 0.5661 | 0.2897 |
110
+ | 0.4819 | 39.45 | 4300 | 0.5663 | 0.2883 |
111
 
112
 
113
  ### Framework versions
114
 
115
  - Transformers 4.37.0.dev0
116
+ - Pytorch 2.1.0+cu121
117
  - Datasets 2.15.0
118
  - Tokenizers 0.15.0
adapter.vie.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:3bc1c4508db41ba0d430ce70644b8005d2b7eed9d9ea3a91567dda801f7c0b8f
3
  size 9203360
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:609f96d3ccb10717479271609e4af446794c133ea252876cf249028d2de84009
3
  size 9203360
model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:01e4410cca7b465c7e6f0af0e692a2d26800130821d1ea0bb1745be5dcccfca5
3
  size 3859295720
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e6f97093159c1acfcc8c943b9280468fed8d2dae617ff9dfb395edd7bf128ce6
3
  size 3859295720