Hugging Face's logo

MelanieKoe
/
w2v2-base-pretrained_lr5e-5_at0.8_da0.9

Automatic Speech Recognition
Transformers
Safetensors
wav2vec2
Generated from Trainer
Inference Endpoints
Model card Files Community
MelanieKoe commited on
Commit
fd677da
•
1 Parent(s): d032801

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: 1.8482
21
- - Wer: 0.1777
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
- | 15.2605 | 6.1 | 250 | 3.7475 | 1.0 |
55
- | 3.1653 | 12.2 | 500 | 3.1554 | 1.0 |
56
- | 3.0106 | 18.29 | 750 | 2.7843 | 1.0 |
57
- | 0.9917 | 24.39 | 1000 | 0.8127 | 0.3319 |
58
- | 0.1612 | 30.49 | 1250 | 1.2476 | 0.2687 |
59
- | 0.0974 | 36.59 | 1500 | 1.4311 | 0.2431 |
60
- | 0.0672 | 42.68 | 1750 | 1.5096 | 0.2008 |
61
- | 0.0505 | 48.78 | 2000 | 1.6350 | 0.1845 |
62
- | 0.0423 | 54.88 | 2250 | 1.7272 | 0.1858 |
63
- | 0.0334 | 60.98 | 2500 | 1.7594 | 0.1833 |
64
- | 0.0306 | 67.07 | 2750 | 1.6631 | 0.1948 |
65
- | 0.0265 | 73.17 | 3000 | 1.8532 | 0.1824 |
66
- | 0.0222 | 79.27 | 3250 | 1.9459 | 0.1845 |
67
- | 0.0201 | 85.37 | 3500 | 1.8337 | 0.1790 |
68
- | 0.0177 | 91.46 | 3750 | 1.8381 | 0.1773 |
69
- | 0.0164 | 97.56 | 4000 | 1.8482 | 0.1777 |
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: 2.3151
21
+ - Wer: 0.1807
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
+ | 18.3155 | 6.1 | 250 | 3.8850 | 1.0 |
55
+ | 3.2773 | 12.2 | 500 | 3.1874 | 1.0 |
56
+ | 3.0747 | 18.29 | 750 | 3.0519 | 1.0 |
57
+ | 1.8747 | 24.39 | 1000 | 1.0181 | 0.5229 |
58
+ | 0.2774 | 30.49 | 1250 | 1.5081 | 0.2191 |
59
+ | 0.1428 | 36.59 | 1500 | 1.5603 | 0.2067 |
60
+ | 0.0989 | 42.68 | 1750 | 1.6293 | 0.1999 |
61
+ | 0.0748 | 48.78 | 2000 | 1.9993 | 0.1978 |
62
+ | 0.0584 | 54.88 | 2250 | 2.0658 | 0.1914 |
63
+ | 0.0481 | 60.98 | 2500 | 2.1470 | 0.1845 |
64
+ | 0.0433 | 67.07 | 2750 | 2.2501 | 0.1833 |
65
+ | 0.038 | 73.17 | 3000 | 2.1957 | 0.1892 |
66
+ | 0.0331 | 79.27 | 3250 | 2.2906 | 0.1845 |
67
+ | 0.0314 | 85.37 | 3500 | 2.3224 | 0.1841 |
68
+ | 0.0279 | 91.46 | 3750 | 2.3041 | 0.1828 |
69
+ | 0.0267 | 97.56 | 4000 | 2.3151 | 0.1807 |
70
 
71
 
72
  ### Framework versions
model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:2e38b724a2b2f01e2c1e81aea736324266ce8cec588210a8a3c98be3319f1d74
3
  size 377629528
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:41c547201aeb28c7cb59a538e21da2c5ce0e8edb1330d3edb08680259f0a0e9a
3
  size 377629528