Hugging Face's logo

deepdml
/
whisper-tiny-ar-mix-norm

Automatic Speech Recognition
Transformers
TensorBoard
Safetensors
Arabic
whisper
Generated from Trainer
Eval Results (legacy)
Model card Files
xet
Metrics Community
deepdml commited on
Commit
cb0ad11
·
verified ·
1 Parent(s): 4a7e03e

End of training

Browse files
Files changed (2) hide show
  1. README.md +28 -39
  2. model.safetensors +1 -1
README.md CHANGED
@@ -7,11 +7,12 @@ base_model: openai/whisper-tiny
7
  tags:
8
  - generated_from_trainer
9
  datasets:
10
- - google/fleurs
11
- - ymoslem/MediaSpeech
12
  - deepdml/Tunisian_MSA
13
- - UBC-NLP/Casablanca
14
  - fixie-ai/common_voice_17_0
 
 
 
 
15
  metrics:
16
  - wer
17
  model-index:
@@ -22,12 +23,13 @@ model-index:
22
  type: automatic-speech-recognition
23
  dataset:
24
  name: Common Voice 17.0
25
- type: google/fleurs
26
  metrics:
27
  - name: Wer
28
  type: wer
29
- value: 52.17678705862912
30
  ---
 
31
  <!-- This model card has been generated automatically according to the information the Trainer had access to. You
32
  should probably proofread and complete it, then remove this comment. -->
33
 
@@ -35,9 +37,9 @@ should probably proofread and complete it, then remove this comment. -->
35
 
36
  This model is a fine-tuned version of [openai/whisper-tiny](https://huggingface.co/openai/whisper-tiny) on the Common Voice 17.0 dataset.
37
  It achieves the following results on the evaluation set:
38
- - Loss: 0.6321
39
- - Wer: 52.1768
40
- - Cer: 18.3597
41
 
42
  ## Model description
43
 
@@ -69,24 +71,24 @@ The following hyperparameters were used during training:
69
 
70
  | Training Loss | Epoch | Step | Validation Loss | Wer | Cer |
71
  |:-------------:|:------:|:-----:|:---------------:|:-------:|:-------:|
72
- | 0.9731 | 0.0556 | 1000 | 0.8246 | 68.5849 | 26.4636 |
73
- | 0.6833 | 0.1111 | 2000 | 0.7503 | 63.3554 | 23.5338 |
74
- | 0.4756 | 0.1667 | 3000 | 0.7112 | 60.5773 | 21.8069 |
75
- | 0.3473 | 0.2222 | 4000 | 0.7019 | 59.5509 | 21.6413 |
76
- | 0.2547 | 0.2778 | 5000 | 0.6910 | 59.1212 | 21.5653 |
77
- | 0.1777 | 0.3333 | 6000 | 0.6924 | 57.6816 | 20.6340 |
78
- | 0.128 | 1.0197 | 7000 | 0.6828 | 57.0996 | 20.5314 |
79
- | 0.11 | 1.0752 | 8000 | 0.6706 | 56.0768 | 20.2707 |
80
- | 0.0869 | 1.1308 | 9000 | 0.6622 | 55.4654 | 20.0036 |
81
- | 0.0714 | 1.1863 | 10000 | 0.6506 | 54.8448 | 19.6163 |
82
- | 0.0594 | 1.2419 | 11000 | 0.6427 | 54.9714 | 19.4470 |
83
- | 0.0541 | 1.2974 | 12000 | 0.6365 | 53.4089 | 19.0258 |
84
- | 0.0484 | 1.353 | 13000 | 0.6371 | 53.7100 | 19.1604 |
85
- | 0.0445 | 2.0393 | 14000 | 0.6359 | 53.8697 | 19.4603 |
86
- | 0.042 | 2.0949 | 15000 | 0.6348 | 52.5403 | 18.6839 |
87
- | 0.0346 | 2.1504 | 16000 | 0.6317 | 52.7809 | 18.6861 |
88
- | 0.0339 | 2.206 | 17000 | 0.6436 | 52.5128 | 18.5581 |
89
- | 0.0404 | 2.2616 | 18000 | 0.6321 | 52.1768 | 18.3597 |
90
 
91
 
92
  ### Framework versions
@@ -95,16 +97,3 @@ The following hyperparameters were used during training:
95
  - Pytorch 2.5.1+cu121
96
  - Datasets 3.6.0
97
  - Tokenizers 0.21.0
98
-
99
- ## Citation
100
-
101
- Please cite the model using the following BibTeX entry:
102
-
103
- ```bibtex
104
- @misc{deepdml/whisper-tiny-ar-mix-norm,
105
- title={Fine-tuned Whisper tiny ASR model for speech recognition in Arabic},
106
- author={Jimenez, David},
107
- howpublished={\url{https://huggingface.co/deepdml/whisper-tiny-ar-mix-norm}},
108
- year={2026}
109
- }
110
- ```
 
7
  tags:
8
  - generated_from_trainer
9
  datasets:
 
 
10
  - deepdml/Tunisian_MSA
 
11
  - fixie-ai/common_voice_17_0
12
+ - pain/MASC
13
+ - google/fleurs
14
+ - UBC-NLP/Casablanca
15
+ - ymoslem/MediaSpeech
16
  metrics:
17
  - wer
18
  model-index:
 
23
  type: automatic-speech-recognition
24
  dataset:
25
  name: Common Voice 17.0
26
+ type: deepdml/Tunisian_MSA
27
  metrics:
28
  - name: Wer
29
  type: wer
30
+ value: 51.49189328143075
31
  ---
32
+
33
  <!-- This model card has been generated automatically according to the information the Trainer had access to. You
34
  should probably proofread and complete it, then remove this comment. -->
35
 
 
37
 
38
  This model is a fine-tuned version of [openai/whisper-tiny](https://huggingface.co/openai/whisper-tiny) on the Common Voice 17.0 dataset.
39
  It achieves the following results on the evaluation set:
40
+ - Loss: 0.5877
41
+ - Wer: 51.4919
42
+ - Cer: 17.9569
43
 
44
  ## Model description
45
 
 
71
 
72
  | Training Loss | Epoch | Step | Validation Loss | Wer | Cer |
73
  |:-------------:|:------:|:-----:|:---------------:|:-------:|:-------:|
74
+ | 1.009 | 0.0556 | 1000 | 0.8091 | 68.4159 | 26.2902 |
75
+ | 0.7599 | 0.1111 | 2000 | 0.7226 | 62.5420 | 23.2768 |
76
+ | 0.562 | 0.1667 | 3000 | 0.6874 | 59.7455 | 21.5321 |
77
+ | 0.4697 | 0.2222 | 4000 | 0.6690 | 57.8891 | 20.7745 |
78
+ | 0.3596 | 0.2778 | 5000 | 0.6601 | 58.0562 | 21.0386 |
79
+ | 0.3156 | 0.3333 | 6000 | 0.6483 | 56.2366 | 20.0099 |
80
+ | 0.2326 | 0.3889 | 7000 | 0.6359 | 55.6215 | 19.8258 |
81
+ | 0.193 | 0.4444 | 8000 | 0.6243 | 55.0118 | 19.7842 |
82
+ | 0.1844 | 0.5 | 9000 | 0.6174 | 55.0100 | 19.8557 |
83
+ | 0.1742 | 0.5556 | 10000 | 0.6142 | 53.9744 | 19.0535 |
84
+ | 0.1408 | 0.6111 | 11000 | 0.6009 | 53.3721 | 19.1199 |
85
+ | 0.1175 | 0.6667 | 12000 | 0.6040 | 52.5679 | 18.4452 |
86
+ | 0.1441 | 0.7222 | 13000 | 0.6043 | 53.1298 | 18.8023 |
87
+ | 0.1245 | 0.7778 | 14000 | 0.5880 | 52.0519 | 18.2682 |
88
+ | 0.1189 | 0.8333 | 15000 | 0.5865 | 51.5103 | 18.1251 |
89
+ | 0.1065 | 0.8889 | 16000 | 0.5863 | 51.3964 | 17.9292 |
90
+ | 0.1266 | 0.9444 | 17000 | 0.5839 | 51.2275 | 17.9617 |
91
+ | 0.1145 | 1.0 | 18000 | 0.5877 | 51.4919 | 17.9569 |
92
 
93
 
94
  ### Framework versions
 
97
  - Pytorch 2.5.1+cu121
98
  - Datasets 3.6.0
99
  - Tokenizers 0.21.0
 
 
 
 
 
 
 
 
 
 
 
 
 
model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:9bd436edbf089e6cbc17820d7ee971ba220860237b01309c6009a6dc410316fe
3
  size 151061672
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:02938f4d787598812f7f7e8df61b5006acae6a488d21ffd7518f062a96a392cd
3
  size 151061672