Hugging Face's logo

MelanieKoe
/
w2v2-base-pretrained_lr5e-5_at0.5_da1

Automatic Speech Recognition
Transformers
Safetensors
wav2vec2
Generated from Trainer
Inference Endpoints
Model card Files Community
MelanieKoe commited on
Commit
8637514
•
1 Parent(s): 260dd7b

Model save

Browse files
Files changed (2) hide show
  1. README.md +19 -19
  2. model.safetensors +1 -1
README.md CHANGED
@@ -17,8 +17,8 @@ should probably proofread and complete it, then remove this comment. -->
17
 
18
  This model is a fine-tuned version of [facebook/wav2vec2-base](https://huggingface.co/facebook/wav2vec2-base) on the None dataset.
19
  It achieves the following results on the evaluation set:
20
- - Loss: 1.3905
21
- - Wer: 0.1747
22
 
23
  ## Model description
24
 
@@ -43,7 +43,7 @@ The following hyperparameters were used during training:
43
  - seed: 42
44
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
45
  - lr_scheduler_type: linear
46
- - lr_scheduler_warmup_steps: 500
47
  - training_steps: 4000
48
  - mixed_precision_training: Native AMP
49
 
@@ -51,22 +51,22 @@ The following hyperparameters were used during training:
51
 
52
  | Training Loss | Epoch | Step | Validation Loss | Wer |
53
  |:-------------:|:-----:|:----:|:---------------:|:------:|
54
- | 15.1994 | 4.17 | 250 | 3.8796 | 1.0 |
55
- | 3.1716 | 8.33 | 500 | 3.1323 | 1.0 |
56
- | 1.9305 | 12.5 | 750 | 0.7720 | 0.5938 |
57
- | 0.3608 | 16.67 | 1000 | 0.7046 | 0.3097 |
58
- | 0.2042 | 20.83 | 1250 | 0.9989 | 0.2422 |
59
- | 0.1228 | 25.0 | 1500 | 1.1311 | 0.1837 |
60
- | 0.0918 | 29.17 | 1750 | 1.1043 | 0.1961 |
61
- | 0.0678 | 33.33 | 2000 | 1.1187 | 0.1850 |
62
- | 0.0532 | 37.5 | 2250 | 1.4067 | 0.1828 |
63
- | 0.0471 | 41.67 | 2500 | 1.2705 | 0.1867 |
64
- | 0.0377 | 45.83 | 2750 | 1.4453 | 0.1790 |
65
- | 0.0336 | 50.0 | 3000 | 1.2754 | 0.1798 |
66
- | 0.0287 | 54.17 | 3250 | 1.3095 | 0.1790 |
67
- | 0.028 | 58.33 | 3500 | 1.3519 | 0.1764 |
68
- | 0.025 | 62.5 | 3750 | 1.4342 | 0.1734 |
69
- | 0.0234 | 66.67 | 4000 | 1.3905 | 0.1747 |
70
 
71
 
72
  ### Framework versions
 
17
 
18
  This model is a fine-tuned version of [facebook/wav2vec2-base](https://huggingface.co/facebook/wav2vec2-base) on the None dataset.
19
  It achieves the following results on the evaluation set:
20
+ - Loss: 1.6472
21
+ - Wer: 0.1786
22
 
23
  ## Model description
24
 
 
43
  - seed: 42
44
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
45
  - lr_scheduler_type: linear
46
+ - lr_scheduler_warmup_steps: 1000
47
  - training_steps: 4000
48
  - mixed_precision_training: Native AMP
49
 
 
51
 
52
  | Training Loss | Epoch | Step | Validation Loss | Wer |
53
  |:-------------:|:-----:|:----:|:---------------:|:------:|
54
+ | 18.9615 | 4.17 | 250 | 4.3427 | 1.0 |
55
+ | 3.3745 | 8.33 | 500 | 3.2218 | 1.0 |
56
+ | 3.1073 | 12.5 | 750 | 3.1350 | 1.0 |
57
+ | 2.1233 | 16.67 | 1000 | 0.8774 | 0.7915 |
58
+ | 0.4649 | 20.83 | 1250 | 0.8879 | 0.3691 |
59
+ | 0.2435 | 25.0 | 1500 | 0.8832 | 0.2119 |
60
+ | 0.1677 | 29.17 | 1750 | 1.2547 | 0.1871 |
61
+ | 0.1148 | 33.33 | 2000 | 1.3004 | 0.1786 |
62
+ | 0.0904 | 37.5 | 2250 | 1.2741 | 0.1995 |
63
+ | 0.0732 | 41.67 | 2500 | 1.4082 | 0.1747 |
64
+ | 0.0613 | 45.83 | 2750 | 1.5234 | 0.1824 |
65
+ | 0.0536 | 50.0 | 3000 | 1.4060 | 0.1798 |
66
+ | 0.046 | 54.17 | 3250 | 1.5669 | 0.1833 |
67
+ | 0.0441 | 58.33 | 3500 | 1.5661 | 0.1807 |
68
+ | 0.0392 | 62.5 | 3750 | 1.6374 | 0.1790 |
69
+ | 0.0384 | 66.67 | 4000 | 1.6472 | 0.1786 |
70
 
71
 
72
  ### Framework versions
model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:c2fd93f316b976791dcf9b866b6fadd8edf030108abb6a08dacfaf2d5288d514
3
  size 377629528
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d9948303cfcf6aa09ae8c8e37c42def688f2498443615b7e642fb8665d38ffcd
3
  size 377629528