Hugging Face's logo

abdusah
/
aradia-ctc-v1

Automatic Speech Recognition
Transformers
PyTorch
wav2vec2
abdusahmbzuai/arabic_speech_massive_300hrs
Generated from Trainer
Inference Endpoints
Model card Files Community
Abdulwahab Sahyoun commited on
Commit
49583eb
•
1 Parent(s): 20df2cf

update model card README.md

Browse files
Files changed (1) hide show
  1. README.md +127 -0
README.md ADDED
@@ -0,0 +1,127 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ tags:
3
+ - generated_from_trainer
4
+ model-index:
5
+ - name: aradia-ctc-v1
6
+ results: []
7
+ ---
8
+
9
+ <!-- This model card has been generated automatically according to the information the Trainer had access to. You
10
+ should probably proofread and complete it, then remove this comment. -->
11
+
12
+ # aradia-ctc-v1
13
+
14
+ This model was trained from scratch on an unknown dataset.
15
+ It achieves the following results on the evaluation set:
16
+ - Loss: 0.7171
17
+ - Wer: 0.3331
18
+
19
+ ## Model description
20
+
21
+ More information needed
22
+
23
+ ## Intended uses & limitations
24
+
25
+ More information needed
26
+
27
+ ## Training and evaluation data
28
+
29
+ More information needed
30
+
31
+ ## Training procedure
32
+
33
+ ### Training hyperparameters
34
+
35
+ The following hyperparameters were used during training:
36
+ - learning_rate: 0.0003
37
+ - train_batch_size: 32
38
+ - eval_batch_size: 32
39
+ - seed: 42
40
+ - gradient_accumulation_steps: 2
41
+ - total_train_batch_size: 64
42
+ - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
43
+ - lr_scheduler_type: linear
44
+ - lr_scheduler_warmup_steps: 500
45
+ - num_epochs: 20.0
46
+ - mixed_precision_training: Native AMP
47
+
48
+ ### Training results
49
+
50
+ | Training Loss | Epoch | Step | Validation Loss | Wer |
51
+ |:-------------:|:-----:|:----:|:---------------:|:------:|
52
+ | No log | 0.22 | 100 | 5.1889 | 1.0 |
53
+ | No log | 0.43 | 200 | 3.1129 | 1.0 |
54
+ | No log | 0.65 | 300 | 3.0503 | 1.0 |
55
+ | No log | 0.87 | 400 | 3.0279 | 1.0 |
56
+ | 6.2756 | 1.09 | 500 | 2.9965 | 1.0 |
57
+ | 6.2756 | 1.3 | 600 | 2.3618 | 0.9993 |
58
+ | 6.2756 | 1.52 | 700 | 1.2715 | 0.8758 |
59
+ | 6.2756 | 1.74 | 800 | 0.9971 | 0.7156 |
60
+ | 6.2756 | 1.96 | 900 | 0.8927 | 0.6382 |
61
+ | 1.712 | 2.17 | 1000 | 0.8252 | 0.5926 |
62
+ | 1.712 | 2.39 | 1100 | 0.7794 | 0.5434 |
63
+ | 1.712 | 2.61 | 1200 | 0.7557 | 0.5092 |
64
+ | 1.712 | 2.83 | 1300 | 0.7347 | 0.5203 |
65
+ | 1.712 | 3.04 | 1400 | 0.7189 | 0.4929 |
66
+ | 0.9305 | 3.26 | 1500 | 0.6820 | 0.4595 |
67
+ | 0.9305 | 3.48 | 1600 | 0.6792 | 0.4504 |
68
+ | 0.9305 | 3.69 | 1700 | 0.6596 | 0.4442 |
69
+ | 0.9305 | 3.91 | 1800 | 0.6756 | 0.4432 |
70
+ | 0.9305 | 4.13 | 1900 | 0.6663 | 0.4392 |
71
+ | 0.737 | 4.35 | 2000 | 0.6479 | 0.4372 |
72
+ | 0.737 | 4.56 | 2100 | 0.6353 | 0.4203 |
73
+ | 0.737 | 4.78 | 2200 | 0.6251 | 0.4088 |
74
+ | 0.737 | 5.0 | 2300 | 0.6209 | 0.4177 |
75
+ | 0.737 | 5.22 | 2400 | 0.6639 | 0.4094 |
76
+ | 0.6247 | 5.43 | 2500 | 0.6408 | 0.3970 |
77
+ | 0.6247 | 5.65 | 2600 | 0.6373 | 0.3932 |
78
+ | 0.6247 | 5.87 | 2700 | 0.6411 | 0.3928 |
79
+ | 0.6247 | 6.09 | 2800 | 0.6378 | 0.3897 |
80
+ | 0.6247 | 6.3 | 2900 | 0.6396 | 0.3929 |
81
+ | 0.5443 | 6.52 | 3000 | 0.6544 | 0.3864 |
82
+ | 0.5443 | 6.74 | 3100 | 0.6218 | 0.3786 |
83
+ | 0.5443 | 6.96 | 3200 | 0.6200 | 0.3784 |
84
+ | 0.5443 | 7.17 | 3300 | 0.6157 | 0.3791 |
85
+ | 0.5443 | 7.39 | 3400 | 0.6317 | 0.3798 |
86
+ | 0.4845 | 7.61 | 3500 | 0.6540 | 0.3771 |
87
+ | 0.4845 | 7.83 | 3600 | 0.6436 | 0.3670 |
88
+ | 0.4845 | 8.04 | 3700 | 0.6335 | 0.3695 |
89
+ | 0.4845 | 8.26 | 3800 | 0.6579 | 0.3610 |
90
+ | 0.4845 | 8.48 | 3900 | 0.6170 | 0.3613 |
91
+ | 0.4279 | 8.69 | 4000 | 0.6523 | 0.3617 |
92
+ | 0.4279 | 8.91 | 4100 | 0.6349 | 0.3577 |
93
+ | 0.4279 | 9.13 | 4200 | 0.6344 | 0.3673 |
94
+ | 0.4279 | 9.35 | 4300 | 0.6215 | 0.3641 |
95
+ | 0.4279 | 9.56 | 4400 | 0.6513 | 0.3608 |
96
+ | 0.3825 | 9.78 | 4500 | 0.6386 | 0.3605 |
97
+ | 0.3825 | 10.0 | 4600 | 0.6724 | 0.3549 |
98
+ | 0.3825 | 10.22 | 4700 | 0.6776 | 0.3602 |
99
+ | 0.3825 | 10.43 | 4800 | 0.6739 | 0.3544 |
100
+ | 0.3825 | 10.65 | 4900 | 0.6688 | 0.3557 |
101
+ | 0.3477 | 10.87 | 5000 | 0.6674 | 0.3564 |
102
+ | 0.3477 | 11.09 | 5100 | 0.6786 | 0.3476 |
103
+ | 0.3477 | 11.3 | 5200 | 0.6818 | 0.3478 |
104
+ | 0.3477 | 11.52 | 5300 | 0.6874 | 0.3470 |
105
+ | 0.3477 | 11.74 | 5400 | 0.6993 | 0.3424 |
106
+ | 0.3101 | 11.96 | 5500 | 0.6950 | 0.3404 |
107
+ | 0.3101 | 12.17 | 5600 | 0.6872 | 0.3406 |
108
+ | 0.3101 | 12.39 | 5700 | 0.6846 | 0.3424 |
109
+ | 0.3101 | 12.61 | 5800 | 0.7051 | 0.3405 |
110
+ | 0.3101 | 12.83 | 5900 | 0.7051 | 0.3378 |
111
+ | 0.2859 | 13.04 | 6000 | 0.6955 | 0.3403 |
112
+ | 0.2859 | 13.26 | 6100 | 0.7115 | 0.3390 |
113
+ | 0.2859 | 13.48 | 6200 | 0.7074 | 0.3384 |
114
+ | 0.2859 | 13.69 | 6300 | 0.7002 | 0.3376 |
115
+ | 0.2859 | 13.91 | 6400 | 0.7171 | 0.3360 |
116
+ | 0.2714 | 14.13 | 6500 | 0.7193 | 0.3341 |
117
+ | 0.2714 | 14.35 | 6600 | 0.7132 | 0.3347 |
118
+ | 0.2714 | 14.56 | 6700 | 0.7184 | 0.3353 |
119
+ | 0.2714 | 14.78 | 6800 | 0.7171 | 0.3331 |
120
+
121
+
122
+ ### Framework versions
123
+
124
+ - Transformers 4.18.0.dev0
125
+ - Pytorch 1.10.2+cu113
126
+ - Datasets 1.18.4
127
+ - Tokenizers 0.11.6