Hugging Face's logo

MelanieKoe
/
w2v2-base-pretrained_lr5e-5_at0.8_da0.1

Automatic Speech Recognition
Transformers
Safetensors
wav2vec2
Generated from Trainer
Inference Endpoints
Model card Files Community
MelanieKoe commited on
Commit
30d4d41
•
1 Parent(s): 857583d

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: 11.1145
21
- - Wer: 0.8466
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
- | 30.6579 | 50.0 | 250 | 9.9946 | 1.0 |
55
- | 6.2256 | 100.0 | 500 | 8.2673 | 1.0 |
56
- | 2.8836 | 150.0 | 750 | 8.1729 | 1.0 |
57
- | 1.2163 | 200.0 | 1000 | 7.8384 | 1.0406 |
58
- | 0.4153 | 250.0 | 1250 | 8.4861 | 0.9833 |
59
- | 0.2165 | 300.0 | 1500 | 8.9662 | 0.9334 |
60
- | 0.1446 | 350.0 | 1750 | 9.3884 | 0.8800 |
61
- | 0.084 | 400.0 | 2000 | 9.8852 | 0.8898 |
62
- | 0.0684 | 450.0 | 2250 | 10.0792 | 0.8988 |
63
- | 0.0512 | 500.0 | 2500 | 10.3480 | 0.8548 |
64
- | 0.0458 | 550.0 | 2750 | 10.6581 | 0.8855 |
65
- | 0.0284 | 600.0 | 3000 | 10.6477 | 0.8800 |
66
- | 0.0198 | 650.0 | 3250 | 10.7117 | 0.8642 |
67
- | 0.0139 | 700.0 | 3500 | 11.0289 | 0.8595 |
68
- | 0.0123 | 750.0 | 3750 | 11.1053 | 0.8432 |
69
- | 0.0089 | 800.0 | 4000 | 11.1145 | 0.8466 |
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: 4.7430
21
+ - Wer: 0.8847
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
+ | 32.6282 | 50.0 | 250 | 4.0044 | 1.0 |
55
+ | 3.5474 | 100.0 | 500 | 3.1597 | 1.0 |
56
+ | 3.0565 | 150.0 | 750 | 3.0680 | 1.0 |
57
+ | 2.8909 | 200.0 | 1000 | 3.0005 | 1.0 |
58
+ | 2.292 | 250.0 | 1250 | 2.6442 | 0.9991 |
59
+ | 1.2329 | 300.0 | 1500 | 2.9696 | 0.9577 |
60
+ | 0.6684 | 350.0 | 1750 | 3.4034 | 0.9453 |
61
+ | 0.4377 | 400.0 | 2000 | 3.7834 | 0.9389 |
62
+ | 0.3292 | 450.0 | 2250 | 3.9480 | 0.9000 |
63
+ | 0.2554 | 500.0 | 2500 | 4.3241 | 0.8906 |
64
+ | 0.2073 | 550.0 | 2750 | 4.4139 | 0.9047 |
65
+ | 0.1762 | 600.0 | 3000 | 4.5423 | 0.9000 |
66
+ | 0.1551 | 650.0 | 3250 | 4.5978 | 0.8817 |
67
+ | 0.1403 | 700.0 | 3500 | 4.6304 | 0.8757 |
68
+ | 0.1286 | 750.0 | 3750 | 4.6839 | 0.8855 |
69
+ | 0.1201 | 800.0 | 4000 | 4.7430 | 0.8847 |
70
 
71
 
72
  ### Framework versions
model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:a1bfc361f6e01b740aa616e21b90ce20c00dd4bb85a46dabc90f5bb38ba01400
3
  size 377629528
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8013ca00df25d94ff6ef4fb1239b51f288e2a905da74e2ffeb65d94e4113b0b2
3
  size 377629528