Hugging Face's logo

maher13
/
arabic-iti

Automatic Speech Recognition
Transformers
PyTorch
wav2vec2
Generated from Trainer
Inference Endpoints
Model card Files Community
maher13 commited on
Commit
efc8558
•
1 Parent(s): 687f779

update model card README.md

Browse files
Files changed (1) hide show
  1. README.md +104 -0
README.md ADDED
@@ -0,0 +1,104 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: apache-2.0
3
+ tags:
4
+ - generated_from_trainer
5
+ datasets:
6
+ - common_voice
7
+ model-index:
8
+ - name: arabic-iti
9
+ results: []
10
+ ---
11
+
12
+ <!-- This model card has been generated automatically according to the information the Trainer had access to. You
13
+ should probably proofread and complete it, then remove this comment. -->
14
+
15
+ # arabic-iti
16
+
17
+ This model is a fine-tuned version of [facebook/wav2vec2-large-xlsr-53](https://huggingface.co/facebook/wav2vec2-large-xlsr-53) on the common_voice dataset.
18
+ It achieves the following results on the evaluation set:
19
+ - Loss: 1.0154
20
+ - Wer: 0.6350
21
+
22
+ ## Model description
23
+
24
+ More information needed
25
+
26
+ ## Intended uses & limitations
27
+
28
+ More information needed
29
+
30
+ ## Training and evaluation data
31
+
32
+ More information needed
33
+
34
+ ## Training procedure
35
+
36
+ ### Training hyperparameters
37
+
38
+ The following hyperparameters were used during training:
39
+ - learning_rate: 0.0005
40
+ - train_batch_size: 8
41
+ - eval_batch_size: 1
42
+ - seed: 42
43
+ - gradient_accumulation_steps: 8
44
+ - total_train_batch_size: 64
45
+ - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
46
+ - lr_scheduler_type: linear
47
+ - lr_scheduler_warmup_steps: 3000
48
+ - num_epochs: 50
49
+ - mixed_precision_training: Native AMP
50
+
51
+ ### Training results
52
+
53
+ | Training Loss | Epoch | Step | Validation Loss | Wer |
54
+ |:-------------:|:-----:|:-----:|:---------------:|:------:|
55
+ | 3.0355 | 2.36 | 400 | 3.0286 | 1.0 |
56
+ | 0.7999 | 4.73 | 800 | 0.8623 | 0.8067 |
57
+ | 0.4485 | 7.1 | 1200 | 0.6920 | 0.6651 |
58
+ | 0.3719 | 9.47 | 1600 | 0.6361 | 0.6591 |
59
+ | 0.3401 | 11.83 | 2000 | 0.6967 | 0.6497 |
60
+ | 0.3222 | 14.2 | 2400 | 0.6697 | 0.6246 |
61
+ | 0.3094 | 16.57 | 2800 | 0.7282 | 0.6537 |
62
+ | 0.2822 | 18.93 | 3200 | 0.8019 | 0.6816 |
63
+ | 0.2446 | 21.3 | 3600 | 0.7622 | 0.6608 |
64
+ | 0.235 | 23.67 | 4000 | 0.8644 | 0.6780 |
65
+ | 0.2362 | 26.04 | 4400 | 0.9083 | 0.6710 |
66
+ | 0.206 | 28.4 | 4800 | 0.8243 | 0.6598 |
67
+ | 0.1765 | 30.77 | 5200 | 0.8614 | 0.6647 |
68
+ | 0.1458 | 33.14 | 5600 | 0.8907 | 0.6447 |
69
+ | 0.1544 | 35.5 | 6000 | 0.9059 | 0.6523 |
70
+ | 0.2402 | 18.88 | 6400 | 0.9639 | 0.6970 |
71
+ | 0.2026 | 20.06 | 6800 | 0.9868 | 0.6817 |
72
+ | 0.185 | 21.24 | 7200 | 1.0043 | 0.6936 |
73
+ | 0.1951 | 22.42 | 7600 | 0.8918 | 0.6795 |
74
+ | 0.1933 | 23.6 | 8000 | 0.9367 | 0.6826 |
75
+ | 0.2272 | 24.78 | 8400 | 0.8540 | 0.6792 |
76
+ | 0.1922 | 25.96 | 8800 | 0.8983 | 0.6657 |
77
+ | 0.1547 | 27.14 | 9200 | 0.9742 | 0.6747 |
78
+ | 0.1579 | 28.32 | 9600 | 0.9066 | 0.6668 |
79
+ | 0.1642 | 29.5 | 10000 | 0.9440 | 0.6790 |
80
+ | 0.1726 | 30.68 | 10400 | 0.9654 | 0.6813 |
81
+ | 0.1656 | 31.86 | 10800 | 0.9880 | 0.6801 |
82
+ | 0.1741 | 33.04 | 11200 | 0.9707 | 0.6584 |
83
+ | 0.1494 | 34.22 | 11600 | 0.9801 | 0.6709 |
84
+ | 0.1482 | 35.4 | 12000 | 0.9258 | 0.6646 |
85
+ | 0.14 | 36.58 | 12400 | 0.9802 | 0.6635 |
86
+ | 0.142 | 37.76 | 12800 | 0.9268 | 0.6524 |
87
+ | 0.1281 | 38.94 | 13200 | 0.9615 | 0.6587 |
88
+ | 0.1051 | 40.12 | 13600 | 0.9721 | 0.6495 |
89
+ | 0.1074 | 41.3 | 14000 | 1.0045 | 0.6582 |
90
+ | 0.0879 | 42.48 | 14400 | 1.0290 | 0.6516 |
91
+ | 0.1015 | 43.66 | 14800 | 1.0514 | 0.6556 |
92
+ | 0.0932 | 44.84 | 15200 | 1.0287 | 0.6450 |
93
+ | 0.1008 | 46.02 | 15600 | 0.9940 | 0.6399 |
94
+ | 0.0968 | 47.2 | 16000 | 1.0206 | 0.6368 |
95
+ | 0.0858 | 48.38 | 16400 | 1.0452 | 0.6361 |
96
+ | 0.0886 | 49.56 | 16800 | 1.0154 | 0.6350 |
97
+
98
+
99
+ ### Framework versions
100
+
101
+ - Transformers 4.11.3
102
+ - Pytorch 1.10.1+cu102
103
+ - Datasets 1.13.3
104
+ - Tokenizers 0.10.3