Hugging Face's logo

nrshoudi
/
wav2vec2-large-xls-r-300m-Arabic-phoneme

Automatic Speech Recognition
Transformers
PyTorch
TensorBoard
wav2vec2
Generated from Trainer
Inference Endpoints
Model card Files Metrics Community
8
nrshoudi commited on
Commit
295366e
•
1 Parent(s): 74c8719

update model card README.md

Browse files
Files changed (1) hide show
  1. README.md +90 -0
README.md ADDED
@@ -0,0 +1,90 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: apache-2.0
3
+ tags:
4
+ - generated_from_trainer
5
+ model-index:
6
+ - name: wav2vec2-large-xls-r-300m-Arabic-phoneme
7
+ results: []
8
+ ---
9
+
10
+ <!-- This model card has been generated automatically according to the information the Trainer had access to. You
11
+ should probably proofread and complete it, then remove this comment. -->
12
+
13
+ # wav2vec2-large-xls-r-300m-Arabic-phoneme
14
+
15
+ This model is a fine-tuned version of [facebook/wav2vec2-large-xlsr-53](https://huggingface.co/facebook/wav2vec2-large-xlsr-53) on the None dataset.
16
+ It achieves the following results on the evaluation set:
17
+ - Loss: 0.0131
18
+ - Cer: 0.0698
19
+
20
+ ## Model description
21
+
22
+ More information needed
23
+
24
+ ## Intended uses & limitations
25
+
26
+ More information needed
27
+
28
+ ## Training and evaluation data
29
+
30
+ More information needed
31
+
32
+ ## Training procedure
33
+
34
+ ### Training hyperparameters
35
+
36
+ The following hyperparameters were used during training:
37
+ - learning_rate: 0.0005
38
+ - train_batch_size: 6
39
+ - eval_batch_size: 8
40
+ - seed: 42
41
+ - gradient_accumulation_steps: 4
42
+ - total_train_batch_size: 24
43
+ - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
44
+ - lr_scheduler_type: linear
45
+ - lr_scheduler_warmup_steps: 250
46
+ - num_epochs: 30
47
+ - mixed_precision_training: Native AMP
48
+
49
+ ### Training results
50
+
51
+ | Training Loss | Epoch | Step | Validation Loss | Cer |
52
+ |:-------------:|:-----:|:----:|:---------------:|:------:|
53
+ | 3.7128 | 1.0 | 136 | 3.4914 | 0.9731 |
54
+ | 2.99 | 2.0 | 272 | 3.2392 | 0.9731 |
55
+ | 1.8777 | 3.0 | 408 | 1.0871 | 0.4466 |
56
+ | 0.5067 | 4.0 | 544 | 0.2068 | 0.1338 |
57
+ | 0.1711 | 5.0 | 680 | 0.0875 | 0.0913 |
58
+ | 0.0905 | 6.0 | 816 | 0.0734 | 0.0843 |
59
+ | 0.0712 | 7.0 | 952 | 0.0433 | 0.0770 |
60
+ | 0.0535 | 8.0 | 1088 | 0.0314 | 0.0733 |
61
+ | 0.0502 | 9.0 | 1224 | 0.0345 | 0.0752 |
62
+ | 0.0501 | 10.0 | 1360 | 0.0265 | 0.0741 |
63
+ | 0.0446 | 11.0 | 1496 | 0.0326 | 0.0741 |
64
+ | 0.0348 | 12.0 | 1632 | 0.0340 | 0.0763 |
65
+ | 0.046 | 13.0 | 1768 | 0.0250 | 0.0735 |
66
+ | 0.0316 | 14.0 | 1904 | 0.0480 | 0.0860 |
67
+ | 0.0198 | 15.0 | 2040 | 0.0267 | 0.0736 |
68
+ | 0.0209 | 16.0 | 2176 | 0.0173 | 0.0713 |
69
+ | 0.0131 | 17.0 | 2312 | 0.0204 | 0.0714 |
70
+ | 0.0184 | 18.0 | 2448 | 0.0183 | 0.0707 |
71
+ | 0.0136 | 19.0 | 2584 | 0.0245 | 0.0717 |
72
+ | 0.0165 | 20.0 | 2720 | 0.0200 | 0.0737 |
73
+ | 0.012 | 21.0 | 2856 | 0.0152 | 0.0703 |
74
+ | 0.0124 | 22.0 | 2992 | 0.0149 | 0.0703 |
75
+ | 0.0118 | 23.0 | 3128 | 0.0168 | 0.0711 |
76
+ | 0.0087 | 24.0 | 3264 | 0.0139 | 0.0702 |
77
+ | 0.0077 | 25.0 | 3400 | 0.0128 | 0.0696 |
78
+ | 0.0103 | 26.0 | 3536 | 0.0121 | 0.0696 |
79
+ | 0.0059 | 27.0 | 3672 | 0.0131 | 0.0700 |
80
+ | 0.0058 | 28.0 | 3808 | 0.0129 | 0.0699 |
81
+ | 0.0067 | 29.0 | 3944 | 0.0132 | 0.0700 |
82
+ | 0.0055 | 30.0 | 4080 | 0.0131 | 0.0698 |
83
+
84
+
85
+ ### Framework versions
86
+
87
+ - Transformers 4.26.1
88
+ - Pytorch 1.13.1+cu116
89
+ - Datasets 2.10.1
90
+ - Tokenizers 0.13.2