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
ef88bf7
·
verified ·
1 Parent(s): d869b2e

Training finished

Browse files
Files changed (5) hide show
  1. README.md +35 -30
  2. generation_config.json +1 -1
  3. model.safetensors +1 -1
  4. runs/Apr27_01-35-42_8696d7960085/events.out.tfevents.1745717745.8696d7960085.19.0 +2 -2
  5. runs/Apr27_01-35-42_8696d7960085/events.out.tfevents.1745749255.8696d7960085.19.1 +3 -0
README.md CHANGED
@@ -21,8 +21,8 @@ should probably proofread and complete it, then remove this comment. -->
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: 0.0032
24
- - Wer: 0.0505
25
- - Cer: 0.0213
26
 
27
  ## Model description
28
 
@@ -50,38 +50,43 @@ The following hyperparameters were used during training:
50
  - optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
51
  - lr_scheduler_type: linear
52
  - lr_scheduler_warmup_steps: 500
53
- - num_epochs: 25
54
  - mixed_precision_training: Native AMP
55
 
56
  ### Training results
57
 
58
- | Training Loss | Epoch | Step | Validation Loss | Wer | Cer |
59
- |:-------------:|:-----:|:----:|:---------------:|:------:|:------:|
60
- | 0.008 | 1.0 | 250 | 0.0033 | 0.0405 | 0.0171 |
61
- | 0.0045 | 2.0 | 500 | 0.0038 | 0.0485 | 0.0201 |
62
- | 0.0011 | 3.0 | 750 | 0.0047 | 0.0510 | 0.0219 |
63
- | 0.0003 | 4.0 | 1000 | 0.0055 | 0.0541 | 0.0220 |
64
- | 0.0004 | 5.0 | 1250 | 0.0056 | 0.0648 | 0.0263 |
65
- | 0.0004 | 6.0 | 1500 | 0.0062 | 0.0648 | 0.0261 |
66
- | 0.0003 | 7.0 | 1750 | 0.0068 | 0.0661 | 0.0254 |
67
- | 0.0003 | 8.0 | 2000 | 0.0073 | 0.0661 | 0.0259 |
68
- | 0.0001 | 9.0 | 2250 | 0.0076 | 0.0673 | 0.0286 |
69
- | 0.0001 | 10.0 | 2500 | 0.0070 | 0.0614 | 0.0250 |
70
- | 0.0001 | 11.0 | 2750 | 0.0075 | 0.0626 | 0.0242 |
71
- | 0.0 | 12.0 | 3000 | 0.0073 | 0.0588 | 0.0237 |
72
- | 0.0 | 13.0 | 3250 | 0.0073 | 0.0595 | 0.0246 |
73
- | 0.0 | 14.0 | 3500 | 0.0072 | 0.0590 | 0.0234 |
74
- | 0.0 | 15.0 | 3750 | 0.0074 | 0.0599 | 0.0239 |
75
- | 0.0 | 16.0 | 4000 | 0.0075 | 0.0608 | 0.0241 |
76
- | 0.0 | 17.0 | 4250 | 0.0076 | 0.0597 | 0.0236 |
77
- | 0.0 | 18.0 | 4500 | 0.0077 | 0.0599 | 0.0233 |
78
- | 0.0 | 19.0 | 4750 | 0.0077 | 0.0597 | 0.0229 |
79
- | 0.0 | 20.0 | 5000 | 0.0078 | 0.0592 | 0.0232 |
80
- | 0.0 | 21.0 | 5250 | 0.0078 | 0.0595 | 0.0229 |
81
- | 0.0 | 22.0 | 5500 | 0.0079 | 0.0595 | 0.0229 |
82
- | 0.0 | 23.0 | 5750 | 0.0079 | 0.0623 | 0.0250 |
83
- | 0.0 | 24.0 | 6000 | 0.0078 | 0.0654 | 0.0271 |
84
- | 0.0 | 25.0 | 6250 | 0.0079 | 0.0623 | 0.0249 |
 
 
 
 
 
85
 
86
 
87
  ### Framework versions
 
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: 0.0032
24
+ - Wer: 0.0519
25
+ - Cer: 0.0222
26
 
27
  ## Model description
28
 
 
50
  - optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
51
  - lr_scheduler_type: linear
52
  - lr_scheduler_warmup_steps: 500
53
+ - num_epochs: 30
54
  - mixed_precision_training: Native AMP
55
 
56
  ### Training results
57
 
58
+ | Training Loss | Epoch | Step | Validation Loss | Wer | Cer |
59
+ |:-------------:|:-----:|:-----:|:---------------:|:------:|:------:|
60
+ | 0.0059 | 1.0 | 375 | 0.0034 | 0.0498 | 0.0226 |
61
+ | 0.0019 | 2.0 | 750 | 0.0046 | 0.0545 | 0.0235 |
62
+ | 0.0013 | 3.0 | 1125 | 0.0056 | 0.0677 | 0.0288 |
63
+ | 0.0008 | 4.0 | 1500 | 0.0062 | 0.0762 | 0.0314 |
64
+ | 0.0006 | 5.0 | 1875 | 0.0065 | 0.0726 | 0.0311 |
65
+ | 0.0005 | 6.0 | 2250 | 0.0063 | 0.0702 | 0.0301 |
66
+ | 0.0004 | 7.0 | 2625 | 0.0075 | 0.0878 | 0.0408 |
67
+ | 0.0004 | 8.0 | 3000 | 0.0072 | 0.0706 | 0.0284 |
68
+ | 0.0005 | 9.0 | 3375 | 0.0074 | 0.0767 | 0.0321 |
69
+ | 0.0003 | 10.0 | 3750 | 0.0074 | 0.0757 | 0.0315 |
70
+ | 0.0003 | 11.0 | 4125 | 0.0077 | 0.0751 | 0.0330 |
71
+ | 0.0001 | 12.0 | 4500 | 0.0076 | 0.0708 | 0.0322 |
72
+ | 0.0002 | 13.0 | 4875 | 0.0082 | 0.0726 | 0.0307 |
73
+ | 0.0001 | 14.0 | 5250 | 0.0081 | 0.0813 | 0.0374 |
74
+ | 0.0002 | 15.0 | 5625 | 0.0085 | 0.0782 | 0.0317 |
75
+ | 0.0002 | 16.0 | 6000 | 0.0090 | 0.0842 | 0.0343 |
76
+ | 0.0 | 17.0 | 6375 | 0.0089 | 0.0804 | 0.0359 |
77
+ | 0.0 | 18.0 | 6750 | 0.0089 | 0.0787 | 0.0347 |
78
+ | 0.0 | 19.0 | 7125 | 0.0094 | 0.0744 | 0.0316 |
79
+ | 0.0001 | 20.0 | 7500 | 0.0100 | 0.0885 | 0.0399 |
80
+ | 0.0 | 21.0 | 7875 | 0.0098 | 0.0795 | 0.0360 |
81
+ | 0.0 | 22.0 | 8250 | 0.0099 | 0.0807 | 0.0370 |
82
+ | 0.0 | 23.0 | 8625 | 0.0100 | 0.0787 | 0.0353 |
83
+ | 0.0 | 24.0 | 9000 | 0.0102 | 0.0796 | 0.0366 |
84
+ | 0.0 | 25.0 | 9375 | 0.0103 | 0.0786 | 0.0358 |
85
+ | 0.0 | 26.0 | 9750 | 0.0103 | 0.0786 | 0.0359 |
86
+ | 0.0 | 27.0 | 10125 | 0.0104 | 0.0784 | 0.0359 |
87
+ | 0.0 | 28.0 | 10500 | 0.0104 | 0.0776 | 0.0355 |
88
+ | 0.0 | 29.0 | 10875 | 0.0100 | 0.0789 | 0.0353 |
89
+ | 0.0 | 30.0 | 11250 | 0.0105 | 0.0773 | 0.0353 |
90
 
91
 
92
  ### Framework versions
generation_config.json CHANGED
@@ -40,7 +40,7 @@
40
  "bos_token_id": 50257,
41
  "decoder_start_token_id": 50258,
42
  "eos_token_id": 50257,
43
- "forced_decoder_ids": [
44
  [
45
  1,
46
  null
 
40
  "bos_token_id": 50257,
41
  "decoder_start_token_id": 50258,
42
  "eos_token_id": 50257,
43
+ "input_ids": [
44
  [
45
  1,
46
  null
model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:ee0ecdc15700706022fcb2fdfbf466a8949a7b4c01e37d6eb77309c8da0f73c2
3
  size 290403936
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:67409be504e7c14986ac3e4e6579637a6718e261ff309a1ad22403f6fd6b2886
3
  size 290403936
runs/Apr27_01-35-42_8696d7960085/events.out.tfevents.1745717745.8696d7960085.19.0 CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:96ae510b8f2b83cf3231026dc0904ef732074a019b5152a9fe67e3b7deac961c
3
- size 111774
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9377093c709017786f2f2678048a3498e12beb64bb459c9c7433475f37c72ada
3
+ size 112128
runs/Apr27_01-35-42_8696d7960085/events.out.tfevents.1745749255.8696d7960085.19.1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:08921860b812a7b55f70f85fecd71b1d713d752bcf2006122a2ca1d869f64f0c
3
+ size 453