Hugging Face's logo

Baybars
/
wav2vec2-xls-r-1b-turkish

Automatic Speech Recognition
Transformers
PyTorch
TensorBoard
Turkish
wav2vec2
common_voice
Generated from Trainer
Inference Endpoints
Model card Files Metrics Community
1
Baybars commited on
Commit
0f5bf70
•
1 Parent(s): 095bf84

update model card README.md

Browse files
Files changed (1) hide show
  1. README.md +30 -36
README.md CHANGED
@@ -1,10 +1,5 @@
1
  ---
2
- language:
3
- - tr
4
- license: apache-2.0
5
  tags:
6
- - automatic-speech-recognition
7
- - common_voice
8
  - generated_from_trainer
9
  datasets:
10
  - common_voice
@@ -18,11 +13,11 @@ should probably proofread and complete it, then remove this comment. -->
18
 
19
  #
20
 
21
- This model is a fine-tuned version of [facebook/wav2vec2-xls-r-1b](https://huggingface.co/facebook/wav2vec2-xls-r-1b) on the COMMON_VOICE - TR dataset.
22
  It achieves the following results on the evaluation set:
23
- - Loss: 0.7314
24
- - Wer: 0.4692
25
- - Cer: 0.1321
26
 
27
  ## Model description
28
 
@@ -41,41 +36,40 @@ More information needed
41
  ### Training hyperparameters
42
 
43
  The following hyperparameters were used during training:
44
- - learning_rate: 0.0005
45
  - train_batch_size: 32
46
  - eval_batch_size: 8
47
  - seed: 42
48
- - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
49
  - lr_scheduler_type: linear
50
- - lr_scheduler_warmup_steps: 500
51
- - num_epochs: 100.0
52
  - mixed_precision_training: Native AMP
53
 
54
  ### Training results
55
 
56
- | Training Loss | Epoch | Step | Validation Loss | Wer | Cer |
57
- |:-------------:|:-----:|:-----:|:---------------:|:------:|:------:|
58
- | 1.0779 | 4.59 | 500 | 0.8260 | 0.7395 | 0.2354 |
59
- | 0.7573 | 9.17 | 1000 | 0.7544 | 0.6960 | 0.2100 |
60
- | 0.8225 | 13.76 | 1500 | 0.6867 | 0.6672 | 0.2021 |
61
- | 0.621 | 18.35 | 2000 | 0.6824 | 0.6209 | 0.1874 |
62
- | 0.6362 | 22.94 | 2500 | 0.6712 | 0.6286 | 0.1904 |
63
- | 0.624 | 27.52 | 3000 | 0.6940 | 0.6116 | 0.1820 |
64
- | 0.4781 | 32.11 | 3500 | 0.6966 | 0.5989 | 0.1735 |
65
- | 0.5685 | 36.7 | 4000 | 0.6742 | 0.5971 | 0.1769 |
66
- | 0.4384 | 41.28 | 4500 | 0.6904 | 0.5999 | 0.1767 |
67
- | 0.5509 | 45.87 | 5000 | 0.6734 | 0.5641 | 0.1692 |
68
- | 0.3665 | 50.46 | 5500 | 0.7018 | 0.5662 | 0.1680 |
69
- | 0.3914 | 55.05 | 6000 | 0.7121 | 0.5552 | 0.1631 |
70
- | 0.2467 | 59.63 | 6500 | 0.6657 | 0.5374 | 0.1563 |
71
- | 0.2576 | 64.22 | 7000 | 0.6920 | 0.5316 | 0.1554 |
72
- | 0.2711 | 68.81 | 7500 | 0.6900 | 0.5176 | 0.1495 |
73
- | 0.2626 | 73.39 | 8000 | 0.6843 | 0.5043 | 0.1454 |
74
- | 0.1377 | 77.98 | 8500 | 0.7383 | 0.5101 | 0.1470 |
75
- | 0.2005 | 82.57 | 9000 | 0.7228 | 0.5045 | 0.1430 |
76
- | 0.1355 | 87.16 | 9500 | 0.7231 | 0.4869 | 0.1375 |
77
- | 0.0431 | 91.74 | 10000 | 0.7397 | 0.4749 | 0.1350 |
78
- | 0.0586 | 96.33 | 10500 | 0.7360 | 0.4754 | 0.1339 |
79
 
80
 
81
  ### Framework versions
 
1
  ---
 
 
 
2
  tags:
 
 
3
  - generated_from_trainer
4
  datasets:
5
  - common_voice
 
13
 
14
  #
15
 
16
+ This model was trained from scratch on the common_voice dataset.
17
  It achieves the following results on the evaluation set:
18
+ - Cer: 0.1339
19
+ - Loss: 0.7360
20
+ - Wer: 0.4754
21
 
22
  ## Model description
23
 
 
36
  ### Training hyperparameters
37
 
38
  The following hyperparameters were used during training:
39
+ - learning_rate: 0.0002
40
  - train_batch_size: 32
41
  - eval_batch_size: 8
42
  - seed: 42
43
+ - optimizer: Adam with betas=(0.999,0.9999) and epsilon=1e-08
44
  - lr_scheduler_type: linear
45
+ - num_epochs: 5.0
 
46
  - mixed_precision_training: Native AMP
47
 
48
  ### Training results
49
 
50
+ | Training Loss | Epoch | Step | Cer | Validation Loss | Wer |
51
+ |:-------------:|:-----:|:-----:|:------:|:---------------:|:------:|
52
+ | 1.0779 | 4.59 | 500 | 0.2354 | 0.8260 | 0.7395 |
53
+ | 0.7573 | 9.17 | 1000 | 0.2100 | 0.7544 | 0.6960 |
54
+ | 0.8225 | 13.76 | 1500 | 0.2021 | 0.6867 | 0.6672 |
55
+ | 0.621 | 18.35 | 2000 | 0.1874 | 0.6824 | 0.6209 |
56
+ | 0.6362 | 22.94 | 2500 | 0.1904 | 0.6712 | 0.6286 |
57
+ | 0.624 | 27.52 | 3000 | 0.1820 | 0.6940 | 0.6116 |
58
+ | 0.4781 | 32.11 | 3500 | 0.1735 | 0.6966 | 0.5989 |
59
+ | 0.5685 | 36.7 | 4000 | 0.1769 | 0.6742 | 0.5971 |
60
+ | 0.4384 | 41.28 | 4500 | 0.1767 | 0.6904 | 0.5999 |
61
+ | 0.5509 | 45.87 | 5000 | 0.1692 | 0.6734 | 0.5641 |
62
+ | 0.3665 | 50.46 | 5500 | 0.1680 | 0.7018 | 0.5662 |
63
+ | 0.3914 | 55.05 | 6000 | 0.1631 | 0.7121 | 0.5552 |
64
+ | 0.2467 | 59.63 | 6500 | 0.1563 | 0.6657 | 0.5374 |
65
+ | 0.2576 | 64.22 | 7000 | 0.1554 | 0.6920 | 0.5316 |
66
+ | 0.2711 | 68.81 | 7500 | 0.1495 | 0.6900 | 0.5176 |
67
+ | 0.2626 | 73.39 | 8000 | 0.1454 | 0.6843 | 0.5043 |
68
+ | 0.1377 | 77.98 | 8500 | 0.1470 | 0.7383 | 0.5101 |
69
+ | 0.2005 | 82.57 | 9000 | 0.1430 | 0.7228 | 0.5045 |
70
+ | 0.1355 | 87.16 | 9500 | 0.1375 | 0.7231 | 0.4869 |
71
+ | 0.0431 | 91.74 | 10000 | 0.1350 | 0.7397 | 0.4749 |
72
+ | 0.0586 | 96.33 | 10500 | 0.1339 | 0.7360 | 0.4754 |
73
 
74
 
75
  ### Framework versions