Hugging Face's logo

adrianSauer
/
wav2vec2-cer-extension

Automatic Speech Recognition
Transformers
TensorBoard
Safetensors
Guaraní
wav2vec2
Generated from Trainer
Inference Endpoints
Model card Files Metrics Community
adrianSauer commited on
Commit
630a7b7
·
verified ·
1 Parent(s): 6c3973c

End of training

Browse files
Files changed (1) hide show
  1. README.md +13 -11
README.md CHANGED
@@ -1,4 +1,5 @@
1
  ---
 
2
  language:
3
  - gn
4
  license: apache-2.0
@@ -19,8 +20,8 @@ should probably proofread and complete it, then remove this comment. -->
19
 
20
  This model is a fine-tuned version of [glob-asr/wav2vec2-large-xls-r-300m-guarani-small](https://huggingface.co/glob-asr/wav2vec2-large-xls-r-300m-guarani-small) on the Common Voice 16 dataset.
21
  It achieves the following results on the evaluation set:
22
- - Loss: 0.3309
23
- - Cer: 7.5608
24
 
25
  ## Model description
26
 
@@ -39,7 +40,7 @@ More information needed
39
  ### Training hyperparameters
40
 
41
  The following hyperparameters were used during training:
42
- - learning_rate: 1e-05
43
  - train_batch_size: 8
44
  - eval_batch_size: 16
45
  - seed: 42
@@ -47,24 +48,25 @@ The following hyperparameters were used during training:
47
  - total_train_batch_size: 16
48
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
49
  - lr_scheduler_type: constant_with_warmup
50
- - lr_scheduler_warmup_steps: 50
51
- - training_steps: 500
52
  - mixed_precision_training: Native AMP
53
 
54
  ### Training results
55
 
56
  | Training Loss | Epoch | Step | Validation Loss | Cer |
57
  |:-------------:|:------:|:----:|:---------------:|:------:|
58
- | 1.3968 | 0.0991 | 100 | 0.3683 | 8.4273 |
59
- | 1.061 | 0.1982 | 200 | 0.3611 | 8.5093 |
60
- | 1.0374 | 0.2973 | 300 | 0.3500 | 8.1463 |
61
- | 0.9825 | 0.3964 | 400 | 0.3458 | 7.9394 |
62
- | 0.9185 | 0.4955 | 500 | 0.3309 | 7.5608 |
 
63
 
64
 
65
  ### Framework versions
66
 
67
- - Transformers 4.44.0
68
  - Pytorch 2.3.1+cu121
69
  - Datasets 2.21.0
70
  - Tokenizers 0.19.1
 
1
  ---
2
+ library_name: transformers
3
  language:
4
  - gn
5
  license: apache-2.0
 
20
 
21
  This model is a fine-tuned version of [glob-asr/wav2vec2-large-xls-r-300m-guarani-small](https://huggingface.co/glob-asr/wav2vec2-large-xls-r-300m-guarani-small) on the Common Voice 16 dataset.
22
  It achieves the following results on the evaluation set:
23
+ - Loss: 0.2425
24
+ - Cer: 5.9170
25
 
26
  ## Model description
27
 
 
40
  ### Training hyperparameters
41
 
42
  The following hyperparameters were used during training:
43
+ - learning_rate: 5e-05
44
  - train_batch_size: 8
45
  - eval_batch_size: 16
46
  - seed: 42
 
48
  - total_train_batch_size: 16
49
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
50
  - lr_scheduler_type: constant_with_warmup
51
+ - lr_scheduler_warmup_steps: 3000
52
+ - training_steps: 3000
53
  - mixed_precision_training: Native AMP
54
 
55
  ### Training results
56
 
57
  | Training Loss | Epoch | Step | Validation Loss | Cer |
58
  |:-------------:|:------:|:----:|:---------------:|:------:|
59
+ | 1.2573 | 0.4955 | 500 | 0.3703 | 8.4904 |
60
+ | 0.9205 | 0.9911 | 1000 | 0.3224 | 7.6296 |
61
+ | 0.7466 | 1.4866 | 1500 | 0.2938 | 7.1221 |
62
+ | 0.6766 | 1.9822 | 2000 | 0.2715 | 6.6510 |
63
+ | 0.5782 | 2.4777 | 2500 | 0.2831 | 7.0497 |
64
+ | 0.5495 | 2.9732 | 3000 | 0.2425 | 5.9170 |
65
 
66
 
67
  ### Framework versions
68
 
69
+ - Transformers 4.44.1
70
  - Pytorch 2.3.1+cu121
71
  - Datasets 2.21.0
72
  - Tokenizers 0.19.1