Hugging Face's logo

MelanieKoe
/
w2v2-base-pretrained_lr5e-5_at0.7_da1

Automatic Speech Recognition
Transformers
Safetensors
wav2vec2
Generated from Trainer
Inference Endpoints
Model card Files Community
MelanieKoe commited on
Commit
7b4abb0
•
1 Parent(s): cac17d5

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: 1.3723
21
- - Wer: 0.1645
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
- | 14.5862 | 4.46 | 250 | 3.6865 | 1.0 |
55
- | 3.1591 | 8.93 | 500 | 3.2812 | 1.0 |
56
- | 2.8713 | 13.39 | 750 | 2.2559 | 0.9658 |
57
- | 0.7672 | 17.86 | 1000 | 0.7625 | 0.3093 |
58
- | 0.216 | 22.32 | 1250 | 0.8197 | 0.2072 |
59
- | 0.1328 | 26.79 | 1500 | 1.0511 | 0.1871 |
60
- | 0.0963 | 31.25 | 1750 | 1.1082 | 0.1803 |
61
- | 0.0808 | 35.71 | 2000 | 1.0651 | 0.1833 |
62
- | 0.0681 | 40.18 | 2250 | 1.1422 | 0.1828 |
63
- | 0.0631 | 44.64 | 2500 | 1.2823 | 0.1726 |
64
- | 0.0573 | 49.11 | 2750 | 1.2592 | 0.1696 |
65
- | 0.0528 | 53.57 | 3000 | 1.2268 | 0.1623 |
66
- | 0.05 | 58.04 | 3250 | 1.2559 | 0.1674 |
67
- | 0.0483 | 62.5 | 3500 | 1.2440 | 0.1700 |
68
- | 0.0426 | 66.96 | 3750 | 1.3886 | 0.1713 |
69
- | 0.0321 | 71.43 | 4000 | 1.3723 | 0.1645 |
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: 1.8380
21
+ - Wer: 0.1653
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
+ | 18.7456 | 4.46 | 250 | 4.0623 | 1.0 |
55
+ | 3.321 | 8.93 | 500 | 3.1987 | 1.0 |
56
+ | 3.0779 | 13.39 | 750 | 3.0954 | 1.0 |
57
+ | 1.8257 | 17.86 | 1000 | 0.8216 | 0.6275 |
58
+ | 0.3449 | 22.32 | 1250 | 1.1757 | 0.3268 |
59
+ | 0.2036 | 26.79 | 1500 | 1.2369 | 0.1982 |
60
+ | 0.1299 | 31.25 | 1750 | 1.1629 | 0.1991 |
61
+ | 0.0998 | 35.71 | 2000 | 1.3491 | 0.1743 |
62
+ | 0.0772 | 40.18 | 2250 | 1.4032 | 0.1730 |
63
+ | 0.0667 | 44.64 | 2500 | 1.7240 | 0.1756 |
64
+ | 0.0558 | 49.11 | 2750 | 1.7005 | 0.1709 |
65
+ | 0.0488 | 53.57 | 3000 | 1.7088 | 0.1717 |
66
+ | 0.0427 | 58.04 | 3250 | 1.6884 | 0.1649 |
67
+ | 0.0389 | 62.5 | 3500 | 1.7467 | 0.1670 |
68
+ | 0.035 | 66.96 | 3750 | 1.8174 | 0.1653 |
69
+ | 0.0332 | 71.43 | 4000 | 1.8380 | 0.1653 |
70
 
71
 
72
  ### Framework versions
model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:8a5714ec7d5f19813ee9d0fab4df22e4286f1e4397bf6c5ecda37ec0182066d9
3
  size 377629528
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f817c83632db769b1209ea4c031dcb9a651af0a23c22eb97caff10984c0768dd
3
  size 377629528