Hugging Face's logo

scanne
/
addison6

Automatic Speech Recognition
Transformers
PyTorch
TensorBoard
wav2vec2
Generated from Trainer
Model card Files
xet
Metrics Community
1
scanne commited on
Commit
5c213c1
·
1 Parent(s): 0b3e2de

update model card README.md

Browse files
Files changed (1) hide show
  1. README.md +30 -30
README.md CHANGED
@@ -14,8 +14,8 @@ should probably proofread and complete it, then remove this comment. -->
14
 
15
  This model is a fine-tuned version of [facebook/wav2vec2-base](https://huggingface.co/facebook/wav2vec2-base) on the None dataset.
16
  It achieves the following results on the evaluation set:
17
- - Loss: 1.7585
18
- - Wer: 0.3809
19
 
20
  ## Model description
21
 
@@ -48,34 +48,34 @@ The following hyperparameters were used during training:
48
 
49
  | Training Loss | Epoch | Step | Validation Loss | Wer |
50
  |:-------------:|:-----:|:-----:|:---------------:|:------:|
51
- | 4.0637 | 1.05 | 500 | 2.9155 | 1.0 |
52
- | 1.8717 | 2.09 | 1000 | 1.3860 | 0.7360 |
53
- | 1.2162 | 3.14 | 1500 | 1.1553 | 0.5856 |
54
- | 1.0079 | 4.18 | 2000 | 1.4409 | 0.5260 |
55
- | 0.8638 | 5.23 | 2500 | 1.0576 | 0.4863 |
56
- | 0.7614 | 6.28 | 3000 | 1.0726 | 0.4846 |
57
- | 0.7057 | 7.32 | 3500 | 1.2451 | 0.4555 |
58
- | 0.6127 | 8.37 | 4000 | 1.2994 | 0.4386 |
59
- | 0.5773 | 9.41 | 4500 | 1.3738 | 0.4495 |
60
- | 0.5394 | 10.46 | 5000 | 1.3360 | 0.4443 |
61
- | 0.509 | 11.51 | 5500 | 1.4233 | 0.4070 |
62
- | 0.4561 | 12.55 | 6000 | 1.4158 | 0.4088 |
63
- | 0.4444 | 13.6 | 6500 | 1.5068 | 0.4110 |
64
- | 0.418 | 14.64 | 7000 | 1.4809 | 0.4146 |
65
- | 0.4034 | 15.69 | 7500 | 1.5531 | 0.4082 |
66
- | 0.3623 | 16.74 | 8000 | 1.5302 | 0.4139 |
67
- | 0.3427 | 17.78 | 8500 | 1.4928 | 0.4055 |
68
- | 0.3325 | 18.83 | 9000 | 1.5403 | 0.4025 |
69
- | 0.3159 | 19.87 | 9500 | 1.6301 | 0.3938 |
70
- | 0.2876 | 20.92 | 10000 | 1.6854 | 0.3985 |
71
- | 0.2752 | 21.97 | 10500 | 1.6375 | 0.3891 |
72
- | 0.2585 | 23.01 | 11000 | 1.6425 | 0.3854 |
73
- | 0.2329 | 24.06 | 11500 | 1.6607 | 0.3907 |
74
- | 0.2242 | 25.1 | 12000 | 1.7301 | 0.3843 |
75
- | 0.2015 | 26.15 | 12500 | 1.7208 | 0.3873 |
76
- | 0.1951 | 27.2 | 13000 | 1.7489 | 0.3779 |
77
- | 0.181 | 28.24 | 13500 | 1.7578 | 0.3816 |
78
- | 0.1728 | 29.29 | 14000 | 1.7585 | 0.3809 |
79
 
80
 
81
  ### Framework versions
 
14
 
15
  This model is a fine-tuned version of [facebook/wav2vec2-base](https://huggingface.co/facebook/wav2vec2-base) on the None dataset.
16
  It achieves the following results on the evaluation set:
17
+ - Loss: 1.8211
18
+ - Wer: 0.3999
19
 
20
  ## Model description
21
 
 
48
 
49
  | Training Loss | Epoch | Step | Validation Loss | Wer |
50
  |:-------------:|:-----:|:-----:|:---------------:|:------:|
51
+ | 4.0858 | 1.04 | 500 | 2.8986 | 0.9999 |
52
+ | 1.8903 | 2.08 | 1000 | 1.3421 | 0.7280 |
53
+ | 1.2374 | 3.12 | 1500 | 1.0214 | 0.6037 |
54
+ | 0.9979 | 4.16 | 2000 | 1.2212 | 0.5556 |
55
+ | 0.8648 | 5.2 | 2500 | 1.2449 | 0.5222 |
56
+ | 0.7663 | 6.24 | 3000 | 1.0319 | 0.4734 |
57
+ | 0.6672 | 7.28 | 3500 | 1.3437 | 0.4786 |
58
+ | 0.6224 | 8.32 | 4000 | 1.3717 | 0.4667 |
59
+ | 0.5785 | 9.36 | 4500 | 1.2714 | 0.4374 |
60
+ | 0.5133 | 10.4 | 5000 | 1.4949 | 0.4508 |
61
+ | 0.4963 | 11.43 | 5500 | 1.3073 | 0.4260 |
62
+ | 0.4455 | 12.47 | 6000 | 1.4153 | 0.4375 |
63
+ | 0.4279 | 13.51 | 6500 | 1.3283 | 0.4164 |
64
+ | 0.415 | 14.55 | 7000 | 1.4219 | 0.4335 |
65
+ | 0.384 | 15.59 | 7500 | 1.5668 | 0.4150 |
66
+ | 0.369 | 16.63 | 8000 | 1.6059 | 0.4096 |
67
+ | 0.3415 | 17.67 | 8500 | 1.4167 | 0.4301 |
68
+ | 0.3358 | 18.71 | 9000 | 1.5522 | 0.4258 |
69
+ | 0.3089 | 19.75 | 9500 | 1.5970 | 0.4185 |
70
+ | 0.2934 | 20.79 | 10000 | 1.7218 | 0.4350 |
71
+ | 0.2714 | 21.83 | 10500 | 1.6608 | 0.4174 |
72
+ | 0.2396 | 22.87 | 11000 | 1.6413 | 0.4100 |
73
+ | 0.2346 | 23.91 | 11500 | 1.7355 | 0.4001 |
74
+ | 0.2171 | 24.95 | 12000 | 1.7991 | 0.3980 |
75
+ | 0.1981 | 25.99 | 12500 | 1.7772 | 0.4037 |
76
+ | 0.1952 | 27.03 | 13000 | 1.7544 | 0.4016 |
77
+ | 0.1812 | 28.07 | 13500 | 1.8241 | 0.4012 |
78
+ | 0.168 | 29.11 | 14000 | 1.8211 | 0.3999 |
79
 
80
 
81
  ### Framework versions