Hugging Face's logo

YoussefAshmawy
/
Graduation_Project_Whisper_base

Automatic Speech Recognition
Transformers
TensorBoard
Safetensors
Arabic
whisper
Generated from Trainer
arabic
quran
Eval Results
Model card Files
xet
Metrics Community
YoussefAshmawy commited on
Commit
e7ac348
·
verified ·
1 Parent(s): b20d13e

Training finished

Browse files
Files changed (3) hide show
  1. README.md +43 -35
  2. model.safetensors +1 -1
  3. runs/Dec12_06-34-07_9326196b5699/events.out.tfevents.1734011864.9326196b5699.23.1 +3 -0
README.md CHANGED
@@ -20,8 +20,8 @@ should probably proofread and complete it, then remove this comment. -->
20
 
21
  This model is a fine-tuned version of [openai/whisper-base](https://huggingface.co/openai/whisper-base) on the quran-ayat-speech-to-text dataset.
22
  It achieves the following results on the evaluation set:
23
- - Loss: 0.0626
24
- - Wer: 0.2836
25
 
26
  ## Model description
27
 
@@ -49,45 +49,53 @@ The following hyperparameters were used during training:
49
  - optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
50
  - lr_scheduler_type: linear
51
  - lr_scheduler_warmup_steps: 500
52
- - num_epochs: 20
53
  - mixed_precision_training: Native AMP
54
 
55
  ### Training results
56
 
57
  | Training Loss | Epoch | Step | Validation Loss | Wer |
58
  |:-------------:|:------:|:----:|:---------------:|:------:|
59
- | 0.0586 | 0.625 | 100 | 0.0596 | 0.3968 |
60
- | 0.0458 | 1.25 | 200 | 0.0517 | 0.3544 |
61
- | 0.048 | 1.875 | 300 | 0.0509 | 0.4300 |
62
- | 0.0382 | 2.5 | 400 | 0.0527 | 0.4262 |
63
- | 0.0317 | 3.125 | 500 | 0.0573 | 0.4164 |
64
- | 0.035 | 3.75 | 600 | 0.0540 | 0.4011 |
65
- | 0.0232 | 4.375 | 700 | 0.0577 | 0.3961 |
66
- | 0.0186 | 5.0 | 800 | 0.0548 | 0.3407 |
67
- | 0.0132 | 5.625 | 900 | 0.0593 | 0.3522 |
68
- | 0.0052 | 6.25 | 1000 | 0.0586 | 0.3284 |
69
- | 0.0071 | 6.875 | 1100 | 0.0583 | 0.2953 |
70
- | 0.0041 | 7.5 | 1200 | 0.0587 | 0.3340 |
71
- | 0.0035 | 8.125 | 1300 | 0.0603 | 0.3633 |
72
- | 0.0026 | 8.75 | 1400 | 0.0616 | 0.3173 |
73
- | 0.0011 | 9.375 | 1500 | 0.0611 | 0.3018 |
74
- | 0.0022 | 10.0 | 1600 | 0.0602 | 0.3343 |
75
- | 0.001 | 10.625 | 1700 | 0.0605 | 0.3049 |
76
- | 0.0004 | 11.25 | 1800 | 0.0610 | 0.2913 |
77
- | 0.0003 | 11.875 | 1900 | 0.0617 | 0.2952 |
78
- | 0.0002 | 12.5 | 2000 | 0.0612 | 0.2991 |
79
- | 0.0002 | 13.125 | 2100 | 0.0630 | 0.2909 |
80
- | 0.0001 | 13.75 | 2200 | 0.0626 | 0.2836 |
81
- | 0.0001 | 14.375 | 2300 | 0.0620 | 0.2804 |
82
- | 0.0001 | 15.0 | 2400 | 0.0622 | 0.2838 |
83
- | 0.0002 | 15.625 | 2500 | 0.0625 | 0.2798 |
84
- | 0.0 | 16.25 | 2600 | 0.0625 | 0.2839 |
85
- | 0.0 | 16.875 | 2700 | 0.0627 | 0.2921 |
86
- | 0.0 | 17.5 | 2800 | 0.0629 | 0.2963 |
87
- | 0.0 | 18.125 | 2900 | 0.0630 | 0.2951 |
88
- | 0.0 | 18.75 | 3000 | 0.0631 | 0.2958 |
89
- | 0.0 | 19.375 | 3100 | 0.0631 | 0.2946 |
90
- | 0.0 | 20.0 | 3200 | 0.0631 | 0.2933 |
 
 
 
 
 
 
 
 
91
 
92
 
93
  ### Framework versions
 
20
 
21
  This model is a fine-tuned version of [openai/whisper-base](https://huggingface.co/openai/whisper-base) on the quran-ayat-speech-to-text dataset.
22
  It achieves the following results on the evaluation set:
23
+ - Loss: 0.0805
24
+ - Wer: 0.2980
25
 
26
  ## Model description
27
 
 
49
  - optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
50
  - lr_scheduler_type: linear
51
  - lr_scheduler_warmup_steps: 500
52
+ - num_epochs: 25
53
  - mixed_precision_training: Native AMP
54
 
55
  ### Training results
56
 
57
  | Training Loss | Epoch | Step | Validation Loss | Wer |
58
  |:-------------:|:------:|:----:|:---------------:|:------:|
59
+ | 0.0604 | 0.625 | 100 | 0.0634 | 0.4079 |
60
+ | 0.0293 | 1.25 | 200 | 0.0594 | 0.3206 |
61
+ | 0.0358 | 1.875 | 300 | 0.0553 | 0.3651 |
62
+ | 0.0215 | 2.5 | 400 | 0.0611 | 0.3698 |
63
+ | 0.0203 | 3.125 | 500 | 0.0712 | 0.5044 |
64
+ | 0.025 | 3.75 | 600 | 0.0674 | 0.4716 |
65
+ | 0.0099 | 4.375 | 700 | 0.0738 | 0.3757 |
66
+ | 0.0118 | 5.0 | 800 | 0.0726 | 0.3868 |
67
+ | 0.0069 | 5.625 | 900 | 0.0738 | 0.3961 |
68
+ | 0.0053 | 6.25 | 1000 | 0.0735 | 0.3327 |
69
+ | 0.0066 | 6.875 | 1100 | 0.0765 | 0.3596 |
70
+ | 0.0026 | 7.5 | 1200 | 0.0755 | 0.3642 |
71
+ | 0.003 | 8.125 | 1300 | 0.0792 | 0.3654 |
72
+ | 0.0023 | 8.75 | 1400 | 0.0779 | 0.3496 |
73
+ | 0.0027 | 9.375 | 1500 | 0.0798 | 0.3407 |
74
+ | 0.0024 | 10.0 | 1600 | 0.0774 | 0.3420 |
75
+ | 0.0008 | 10.625 | 1700 | 0.0767 | 0.3425 |
76
+ | 0.0005 | 11.25 | 1800 | 0.0765 | 0.3533 |
77
+ | 0.0004 | 11.875 | 1900 | 0.0793 | 0.3187 |
78
+ | 0.0002 | 12.5 | 2000 | 0.0775 | 0.3371 |
79
+ | 0.0001 | 13.125 | 2100 | 0.0789 | 0.3268 |
80
+ | 0.0002 | 13.75 | 2200 | 0.0784 | 0.3295 |
81
+ | 0.0 | 14.375 | 2300 | 0.0787 | 0.3142 |
82
+ | 0.0001 | 15.0 | 2400 | 0.0800 | 0.3081 |
83
+ | 0.0001 | 15.625 | 2500 | 0.0795 | 0.3105 |
84
+ | 0.0 | 16.25 | 2600 | 0.0791 | 0.3143 |
85
+ | 0.0 | 16.875 | 2700 | 0.0805 | 0.3158 |
86
+ | 0.0 | 17.5 | 2800 | 0.0806 | 0.3141 |
87
+ | 0.0001 | 18.125 | 2900 | 0.0800 | 0.3014 |
88
+ | 0.0 | 18.75 | 3000 | 0.0798 | 0.3165 |
89
+ | 0.0 | 19.375 | 3100 | 0.0801 | 0.3012 |
90
+ | 0.0 | 20.0 | 3200 | 0.0805 | 0.2980 |
91
+ | 0.0 | 20.625 | 3300 | 0.0807 | 0.2963 |
92
+ | 0.0 | 21.25 | 3400 | 0.0807 | 0.3027 |
93
+ | 0.0 | 21.875 | 3500 | 0.0808 | 0.3020 |
94
+ | 0.0 | 22.5 | 3600 | 0.0809 | 0.3006 |
95
+ | 0.0 | 23.125 | 3700 | 0.0810 | 0.3002 |
96
+ | 0.0 | 23.75 | 3800 | 0.0810 | 0.3055 |
97
+ | 0.0 | 24.375 | 3900 | 0.0810 | 0.3017 |
98
+ | 0.0 | 25.0 | 4000 | 0.0810 | 0.3008 |
99
 
100
 
101
  ### Framework versions
model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:1f4b148b971e58c49c5284c7aa9a12b3a25b26aa7116d308413cc483ad987d36
3
  size 290403936
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:87e27045abeb5c54540015c8f98d1a620e25f0270fbdcc70cf690eb01477e7b7
3
  size 290403936
runs/Dec12_06-34-07_9326196b5699/events.out.tfevents.1734011864.9326196b5699.23.1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1e2b38fac0fccf2f35f83abc4fe0e40f8311ec20f072109f4ee5233f6682363c
3
+ size 406