Hugging Face's logo

EYEDOL
/
whisper-tiny-yoruba

Automatic Speech Recognition
Transformers
Safetensors
Yoruba
whisper
Generated from Trainer
Eval Results (legacy)
Model card Files
xet
 Community
EYEDOL commited on
Commit
39dfa88
·
verified ·
1 Parent(s): f3797ac

Model save

Browse files
Files changed (2) hide show
  1. README.md +21 -33
  2. model.safetensors +1 -1
README.md CHANGED
@@ -1,40 +1,26 @@
1
  ---
2
  library_name: transformers
3
- language:
4
- - yo
5
  license: apache-2.0
6
- base_model: openai/whisper-tiny
7
  tags:
8
  - generated_from_trainer
9
- datasets:
10
- - EYEDOL/naija-voices-yoruba-split_0-0
11
  metrics:
12
  - wer
13
  model-index:
14
- - name: EYEDOL/whisper-tiny-yoruba
15
- results:
16
- - task:
17
- name: Automatic Speech Recognition
18
- type: automatic-speech-recognition
19
- dataset:
20
- name: EYEDOL/naija-voices-yoruba-split_0-0
21
- type: EYEDOL/naija-voices-yoruba-split_0-0
22
- metrics:
23
- - name: Wer
24
- type: wer
25
- value: 0.7941161780130862
26
  ---
27
 
28
  <!-- This model card has been generated automatically according to the information the Trainer had access to. You
29
  should probably proofread and complete it, then remove this comment. -->
30
 
31
- # EYEDOL/whisper-tiny-yoruba
32
 
33
- This model is a fine-tuned version of [openai/whisper-tiny](https://huggingface.co/openai/whisper-tiny) on the EYEDOL/naija-voices-yoruba-split_0-0 dataset.
34
  It achieves the following results on the evaluation set:
35
- - Loss: 0.8766
36
- - Wer Ortho: 0.8499
37
- - Wer: 0.7941
38
 
39
  ## Model description
40
 
@@ -60,23 +46,25 @@ The following hyperparameters were used during training:
60
  - optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
61
  - lr_scheduler_type: constant_with_warmup
62
  - lr_scheduler_warmup_steps: 0.1
63
- - num_epochs: 10
64
  - mixed_precision_training: Native AMP
65
 
66
  ### Training results
67
 
68
  | Training Loss | Epoch | Step | Validation Loss | Wer Ortho | Wer |
69
  |:-------------:|:-----:|:----:|:---------------:|:---------:|:------:|
70
- | 5.2951 | 1.0 | 583 | 1.3824 | 0.9785 | 0.9642 |
71
- | 2.3739 | 2.0 | 1166 | 1.1176 | 0.9127 | 0.9391 |
72
- | 1.9820 | 3.0 | 1749 | 1.0172 | 0.9093 | 0.8689 |
73
- | 1.7693 | 4.0 | 2332 | 0.9679 | 0.8792 | 0.8545 |
74
- | 1.6164 | 5.0 | 2915 | 0.9294 | 0.9233 | 0.8784 |
75
- | 1.4949 | 6.0 | 3498 | 0.9078 | 0.8681 | 0.7986 |
76
- | 1.3927 | 7.0 | 4081 | 0.8912 | 0.8216 | 0.7534 |
77
- | 1.3025 | 8.0 | 4664 | 0.8809 | 0.8740 | 0.7988 |
78
- | 1.2197 | 9.0 | 5247 | 0.8761 | 0.8469 | 0.7709 |
79
- | 1.1436 | 10.0 | 5830 | 0.8766 | 0.8499 | 0.7941 |
 
 
80
 
81
 
82
  ### Framework versions
 
1
  ---
2
  library_name: transformers
 
 
3
  license: apache-2.0
4
+ base_model: EYEDOL/whisper-tiny-yoruba
5
  tags:
6
  - generated_from_trainer
 
 
7
  metrics:
8
  - wer
9
  model-index:
10
+ - name: whisper-tiny-yoruba
11
+ results: []
 
 
 
 
 
 
 
 
 
 
12
  ---
13
 
14
  <!-- This model card has been generated automatically according to the information the Trainer had access to. You
15
  should probably proofread and complete it, then remove this comment. -->
16
 
17
+ # whisper-tiny-yoruba
18
 
19
+ This model is a fine-tuned version of [EYEDOL/whisper-tiny-yoruba](https://huggingface.co/EYEDOL/whisper-tiny-yoruba) on an unknown dataset.
20
  It achieves the following results on the evaluation set:
21
+ - Loss: 0.8526
22
+ - Wer Ortho: 0.7773
23
+ - Wer: 0.7006
24
 
25
  ## Model description
26
 
 
46
  - optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
47
  - lr_scheduler_type: constant_with_warmup
48
  - lr_scheduler_warmup_steps: 0.1
49
+ - num_epochs: 12
50
  - mixed_precision_training: Native AMP
51
 
52
  ### Training results
53
 
54
  | Training Loss | Epoch | Step | Validation Loss | Wer Ortho | Wer |
55
  |:-------------:|:-----:|:----:|:---------------:|:---------:|:------:|
56
+ | 1.7353 | 1.0 | 583 | 0.8680 | 0.9448 | 0.8654 |
57
+ | 1.6436 | 2.0 | 1166 | 0.8494 | 0.8213 | 0.7490 |
58
+ | 1.5173 | 3.0 | 1749 | 0.8319 | 0.8237 | 0.7470 |
59
+ | 1.4143 | 4.0 | 2332 | 0.8215 | 0.7845 | 0.7128 |
60
+ | 1.3252 | 5.0 | 2915 | 0.8135 | 0.8788 | 0.7910 |
61
+ | 1.2425 | 6.0 | 3498 | 0.8106 | 0.7988 | 0.7224 |
62
+ | 1.1664 | 7.0 | 4081 | 0.8118 | 0.8508 | 0.7635 |
63
+ | 1.0950 | 8.0 | 4664 | 0.8156 | 0.7628 | 0.6813 |
64
+ | 1.0273 | 9.0 | 5247 | 0.8191 | 0.7867 | 0.7204 |
65
+ | 0.9611 | 10.0 | 5830 | 0.8292 | 0.7736 | 0.6948 |
66
+ | 0.8975 | 11.0 | 6413 | 0.8353 | 0.8007 | 0.7126 |
67
+ | 0.8363 | 12.0 | 6996 | 0.8526 | 0.7773 | 0.7006 |
68
 
69
 
70
  ### Framework versions
model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:fccec7edc19101eedfb2926e2ac1ec0cdd7e7ee9d103dedf1003e1f0f3c0a6c4
3
  size 151061672
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3d65b17d1bf04cd60b0a55b659d28133fda045bc1e9809d07e829d00d8426cef
3
  size 151061672