Hugging Face's logo

Taeyeun72
/
whisper-small-denoising_2

Automatic Speech Recognition
Transformers
PyTorch
Korean
whisper
hf-asr-leaderboard
Generated from Trainer
Eval Results
Inference Endpoints
Model card Files Community
Taeyeun72 commited on
Commit
2f82f7e
·
1 Parent(s): e13541a

End of training

Browse files
Files changed (2) hide show
  1. README.md +25 -25
  2. pytorch_model.bin +1 -1
README.md CHANGED
@@ -25,7 +25,7 @@ model-index:
25
  metrics:
26
  - name: Wer
27
  type: wer
28
- value: 23.17175679462034
29
  ---
30
 
31
  <!-- This model card has been generated automatically according to the information the Trainer had access to. You
@@ -35,9 +35,9 @@ should probably proofread and complete it, then remove this comment. -->
35
 
36
  This model is a fine-tuned version of [openai/whisper-small](https://huggingface.co/openai/whisper-small) on the whisper-kor3_de_2 dataset.
37
  It achieves the following results on the evaluation set:
38
- - Loss: 0.3511
39
- - Wer: 23.1718
40
- - Cer: 10.7478
41
 
42
  ## Model description
43
 
@@ -62,33 +62,33 @@ The following hyperparameters were used during training:
62
  - seed: 42
63
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
64
  - lr_scheduler_type: linear
65
- - lr_scheduler_warmup_steps: 100
66
  - training_steps: 1000
67
 
68
  ### Training results
69
 
70
  | Training Loss | Epoch | Step | Validation Loss | Wer | Cer |
71
  |:-------------:|:-----:|:----:|:---------------:|:-------:|:-------:|
72
- | 0.2928 | 0.21 | 50 | 0.3210 | 23.0877 | 11.0450 |
73
- | 0.2778 | 0.42 | 100 | 0.3221 | 22.6394 | 10.6106 |
74
- | 0.267 | 0.64 | 150 | 0.3227 | 23.0597 | 10.8621 |
75
- | 0.2828 | 0.85 | 200 | 0.3240 | 23.2558 | 10.7859 |
76
- | 0.2293 | 1.06 | 250 | 0.3178 | 22.4152 | 10.3819 |
77
- | 0.156 | 1.27 | 300 | 0.3244 | 23.6761 | 10.9612 |
78
- | 0.156 | 1.48 | 350 | 0.3244 | 24.2925 | 12.2647 |
79
- | 0.1652 | 1.69 | 400 | 0.3223 | 22.7515 | 10.6563 |
80
- | 0.1525 | 1.91 | 450 | 0.3238 | 23.1998 | 10.8011 |
81
- | 0.0744 | 2.12 | 500 | 0.3267 | 22.8075 | 10.6563 |
82
- | 0.0847 | 2.33 | 550 | 0.3277 | 22.6394 | 10.7935 |
83
- | 0.0772 | 2.54 | 600 | 0.3304 | 22.7515 | 10.6106 |
84
- | 0.086 | 2.75 | 650 | 0.3324 | 22.8635 | 10.4962 |
85
- | 0.0827 | 2.97 | 700 | 0.3314 | 23.0877 | 10.8469 |
86
- | 0.0431 | 3.18 | 750 | 0.3414 | 22.9756 | 10.8164 |
87
- | 0.043 | 3.39 | 800 | 0.3438 | 22.6674 | 10.6411 |
88
- | 0.0384 | 3.6 | 850 | 0.3450 | 22.8355 | 10.6487 |
89
- | 0.0409 | 3.81 | 900 | 0.3487 | 23.0317 | 10.6106 |
90
- | 0.0348 | 4.03 | 950 | 0.3481 | 23.1718 | 10.7478 |
91
- | 0.0304 | 4.24 | 1000 | 0.3511 | 23.1718 | 10.7478 |
92
 
93
 
94
  ### Framework versions
 
25
  metrics:
26
  - name: Wer
27
  type: wer
28
+ value: 23.271767810026386
29
  ---
30
 
31
  <!-- This model card has been generated automatically according to the information the Trainer had access to. You
 
35
 
36
  This model is a fine-tuned version of [openai/whisper-small](https://huggingface.co/openai/whisper-small) on the whisper-kor3_de_2 dataset.
37
  It achieves the following results on the evaluation set:
38
+ - Loss: 0.3685
39
+ - Wer: 23.2718
40
+ - Cer: 10.5941
41
 
42
  ## Model description
43
 
 
62
  - seed: 42
63
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
64
  - lr_scheduler_type: linear
65
+ - lr_scheduler_warmup_steps: 200
66
  - training_steps: 1000
67
 
68
  ### Training results
69
 
70
  | Training Loss | Epoch | Step | Validation Loss | Wer | Cer |
71
  |:-------------:|:-----:|:----:|:---------------:|:-------:|:-------:|
72
+ | 0.3178 | 0.21 | 50 | 0.3333 | 23.0343 | 11.1119 |
73
+ | 0.2684 | 0.42 | 100 | 0.3293 | 22.9024 | 10.6660 |
74
+ | 0.279 | 0.64 | 150 | 0.3264 | 23.1662 | 10.8458 |
75
+ | 0.2813 | 0.85 | 200 | 0.3314 | 28.3905 | 14.9597 |
76
+ | 0.2363 | 1.06 | 250 | 0.3325 | 31.3720 | 16.7002 |
77
+ | 0.1909 | 1.27 | 300 | 0.3333 | 26.4908 | 13.0538 |
78
+ | 0.171 | 1.48 | 350 | 0.3384 | 24.4063 | 12.1044 |
79
+ | 0.1699 | 1.69 | 400 | 0.3330 | 22.6649 | 10.3711 |
80
+ | 0.1824 | 1.91 | 450 | 0.3336 | 23.3509 | 10.8242 |
81
+ | 0.0784 | 2.12 | 500 | 0.3425 | 22.7441 | 10.4574 |
82
+ | 0.0694 | 2.33 | 550 | 0.3492 | 23.2982 | 10.7379 |
83
+ | 0.0946 | 2.54 | 600 | 0.3442 | 24.2744 | 11.4212 |
84
+ | 0.0785 | 2.75 | 650 | 0.3486 | 22.6913 | 10.4646 |
85
+ | 0.0838 | 2.97 | 700 | 0.3466 | 22.7441 | 10.5941 |
86
+ | 0.0423 | 3.18 | 750 | 0.3600 | 24.2480 | 11.3780 |
87
+ | 0.0448 | 3.39 | 800 | 0.3615 | 23.0871 | 10.5509 |
88
+ | 0.0492 | 3.6 | 850 | 0.3640 | 23.3509 | 10.6228 |
89
+ | 0.04 | 3.81 | 900 | 0.3649 | 23.4565 | 10.5941 |
90
+ | 0.0385 | 4.03 | 950 | 0.3635 | 24.1689 | 11.2558 |
91
+ | 0.0276 | 4.24 | 1000 | 0.3685 | 23.2718 | 10.5941 |
92
 
93
 
94
  ### Framework versions
pytorch_model.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:7aa404f477c6b597653915c54cbe50e9f262d3ed5653e1d8dbdff6dee09aaeb5
3
  size 967102729
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2bbbe808498f320f32e75f27c094402e92177a8bdb86af35ed08df18bb0212c5
3
  size 967102729