deepdml commited on
Commit
648323e
·
verified ·
1 Parent(s): a3aa927

End of training

Browse files
Files changed (2) hide show
  1. README.md +19 -30
  2. model.safetensors +1 -1
README.md CHANGED
@@ -6,10 +6,11 @@ base_model: openai/whisper-tiny
6
  tags:
7
  - generated_from_trainer
8
  datasets:
 
9
  - google/fleurs
10
- - deepdml/iisc-mile-tamil-asr
11
  - fixie-ai/common_voice_17_0
12
- - deepdml/microsoft-speech-corpus-indian
 
13
  metrics:
14
  - wer
15
  model-index:
@@ -20,12 +21,13 @@ model-index:
20
  type: automatic-speech-recognition
21
  dataset:
22
  name: Common Voice 17.0
23
- type: google/fleurs
24
  metrics:
25
  - name: Wer
26
  type: wer
27
- value: 51.23597531913797
28
  ---
 
29
  <!-- This model card has been generated automatically according to the information the Trainer had access to. You
30
  should probably proofread and complete it, then remove this comment. -->
31
 
@@ -33,9 +35,9 @@ should probably proofread and complete it, then remove this comment. -->
33
 
34
  This model is a fine-tuned version of [openai/whisper-tiny](https://huggingface.co/openai/whisper-tiny) on the Common Voice 17.0 dataset.
35
  It achieves the following results on the evaluation set:
36
- - Loss: 0.2638
37
- - Wer: 51.2360
38
- - Cer: 11.6333
39
 
40
  ## Model description
41
 
@@ -65,16 +67,16 @@ The following hyperparameters were used during training:
65
 
66
  ### Training results
67
 
68
- | Training Loss | Epoch | Step | Validation Loss | Wer | Cer |
69
- |:-------------:|:------:|:----:|:---------------:|:-------:|:-------:|
70
- | 0.2297 | 0.125 | 1000 | 0.3368 | 61.1882 | 15.1244 |
71
- | 0.151 | 0.25 | 2000 | 0.3126 | 57.7102 | 13.8047 |
72
- | 0.1576 | 0.375 | 3000 | 0.2900 | 55.1687 | 13.0734 |
73
- | 0.0818 | 0.5 | 4000 | 0.2829 | 53.7672 | 12.5394 |
74
- | 0.1168 | 0.625 | 5000 | 0.2713 | 53.2661 | 12.2708 |
75
- | 0.0752 | 0.75 | 6000 | 0.2706 | 52.0746 | 11.7892 |
76
- | 0.1097 | 0.875 | 7000 | 0.2647 | 51.7925 | 11.8749 |
77
- | 0.0836 | 1.0961 | 8000 | 0.2638 | 51.2360 | 11.6333 |
78
 
79
 
80
  ### Framework versions
@@ -83,16 +85,3 @@ The following hyperparameters were used during training:
83
  - Pytorch 2.3.0+cu121
84
  - Datasets 2.19.1
85
  - Tokenizers 0.19.1
86
-
87
- ## Citation
88
-
89
- Please cite the model using the following BibTeX entry:
90
-
91
- ```bibtex
92
- @misc{deepdml/whisper-tiny-ta-mix-norm,
93
- title={Fine-tuned Whisper tiny ASR model for speech recognition in Tamil},
94
- author={Jimenez, David},
95
- howpublished={\url{https://huggingface.co/deepdml/whisper-tiny-ta-mix-norm}},
96
- year={2026}
97
- }
98
- ```
 
6
  tags:
7
  - generated_from_trainer
8
  datasets:
9
+ - deepdml/microsoft-speech-corpus-indian
10
  - google/fleurs
 
11
  - fixie-ai/common_voice_17_0
12
+ - ai4bharat/Kathbath
13
+ - deepdml/iisc-mile-tamil-asr
14
  metrics:
15
  - wer
16
  model-index:
 
21
  type: automatic-speech-recognition
22
  dataset:
23
  name: Common Voice 17.0
24
+ type: deepdml/microsoft-speech-corpus-indian
25
  metrics:
26
  - name: Wer
27
  type: wer
28
+ value: 50.94614264919942
29
  ---
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
 
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: 0.2583
39
+ - Wer: 50.9461
40
+ - Cer: 11.6470
41
 
42
  ## Model description
43
 
 
67
 
68
  ### Training results
69
 
70
+ | Training Loss | Epoch | Step | Validation Loss | Wer | Cer |
71
+ |:-------------:|:-----:|:----:|:---------------:|:-------:|:-------:|
72
+ | 0.2315 | 0.125 | 1000 | 0.3384 | 62.1530 | 15.6705 |
73
+ | 0.1623 | 0.25 | 2000 | 0.2973 | 56.7273 | 13.4770 |
74
+ | 0.1716 | 0.375 | 3000 | 0.2856 | 55.2872 | 13.0946 |
75
+ | 0.1572 | 0.5 | 4000 | 0.2676 | 52.6516 | 12.2552 |
76
+ | 0.1475 | 0.625 | 5000 | 0.2650 | 52.0655 | 12.0396 |
77
+ | 0.1656 | 0.75 | 6000 | 0.2610 | 51.5322 | 11.9197 |
78
+ | 0.1048 | 0.875 | 7000 | 0.2561 | 50.7993 | 11.4955 |
79
+ | 0.1166 | 1.0 | 8000 | 0.2583 | 50.9461 | 11.6470 |
80
 
81
 
82
  ### Framework versions
 
85
  - Pytorch 2.3.0+cu121
86
  - Datasets 2.19.1
87
  - Tokenizers 0.19.1
 
 
 
 
 
 
 
 
 
 
 
 
 
model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:42fbdd6bd9b1d8b875600d238924636b7c44f2d1b290fef0e565ce482bbdb5e2
3
  size 151061672
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:79f57b39a899c264ade9bb215625da3d93e4ebeda72fc7ff68d1b3a76f7065ee
3
  size 151061672