Hugging Face's logo

adrianSauer
/
wav2vec2-cer

Automatic Speech Recognition
Transformers
TensorBoard
Safetensors
Guaraní
wav2vec2
Generated from Trainer
Model card Files Metrics Community
adrianSauer commited on
Commit
71ba1be
·
verified ·
1 Parent(s): 760c568

End of training

Browse files
Files changed (1) hide show
  1. README.md +15 -15
README.md CHANGED
@@ -8,19 +8,19 @@ tags:
8
  datasets:
9
  - mozilla-foundation/common_voice_16_1
10
  model-index:
11
- - name: Common Voice 16-10
12
  results: []
13
  ---
14
 
15
  <!-- This model card has been generated automatically according to the information the Trainer had access to. You
16
  should probably proofread and complete it, then remove this comment. -->
17
 
18
- # Common Voice 16-10
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: 0.3185
23
- - Cer: 7.1509
24
 
25
  ## Model description
26
 
@@ -45,24 +45,24 @@ The following hyperparameters were used during training:
45
  - seed: 42
46
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
47
  - lr_scheduler_type: constant_with_warmup
48
- - lr_scheduler_warmup_steps: 1
49
  - training_steps: 500
50
  - mixed_precision_training: Native AMP
51
 
52
  ### Training results
53
 
54
- | Training Loss | Epoch | Step | Validation Loss | Cer |
55
- |:-------------:|:-----:|:----:|:---------------:|:------:|
56
- | 0.3978 | 1.01 | 100 | 0.3420 | 8.0058 |
57
- | 0.3263 | 2.02 | 200 | 0.3318 | 7.7482 |
58
- | 0.28 | 3.03 | 300 | 0.3226 | 7.4710 |
59
- | 0.2505 | 4.04 | 400 | 0.3173 | 7.2758 |
60
- | 0.2427 | 5.05 | 500 | 0.3185 | 7.1509 |
61
 
62
 
63
  ### Framework versions
64
 
65
- - Transformers 4.38.2
66
  - Pytorch 2.2.1+cu121
67
- - Datasets 2.18.0
68
- - Tokenizers 0.15.2
 
8
  datasets:
9
  - mozilla-foundation/common_voice_16_1
10
  model-index:
11
+ - name: Common Voice 16
12
  results: []
13
  ---
14
 
15
  <!-- This model card has been generated automatically according to the information the Trainer had access to. You
16
  should probably proofread and complete it, then remove this comment. -->
17
 
18
+ # Common Voice 16
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: 0.3202
23
+ - Cer: 7.2954
24
 
25
  ## Model description
26
 
 
45
  - seed: 42
46
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
47
  - lr_scheduler_type: constant_with_warmup
48
+ - lr_scheduler_warmup_steps: 50
49
  - training_steps: 500
50
  - mixed_precision_training: Native AMP
51
 
52
  ### Training results
53
 
54
+ | Training Loss | Epoch | Step | Validation Loss | Cer |
55
+ |:-------------:|:------:|:----:|:---------------:|:------:|
56
+ | 0.4174 | 1.0101 | 100 | 0.3535 | 8.1385 |
57
+ | 0.3411 | 2.0202 | 200 | 0.3387 | 7.8574 |
58
+ | 0.2905 | 3.0303 | 300 | 0.3278 | 7.6076 |
59
+ | 0.2591 | 4.0404 | 400 | 0.3214 | 7.3734 |
60
+ | 0.251 | 5.0505 | 500 | 0.3202 | 7.2954 |
61
 
62
 
63
  ### Framework versions
64
 
65
+ - Transformers 4.40.1
66
  - Pytorch 2.2.1+cu121
67
+ - Datasets 2.19.0
68
+ - Tokenizers 0.19.1