Hugging Face's logo

abdusah
/
aradia-ctc-data2vec-ft

Automatic Speech Recognition
Transformers
PyTorch
data2vec-audio
abdusahmbzuai/arabic_speech_massive_300hrs
Generated from Trainer
Inference Endpoints
Model card Files Community
Abdulwahab Sahyoun commited on
Commit
63d6951
•
1 Parent(s): 2b31db8

update model card README.md

Browse files
Files changed (1) hide show
  1. README.md +128 -0
README.md ADDED
@@ -0,0 +1,128 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ tags:
3
+ - generated_from_trainer
4
+ model-index:
5
+ - name: aradia-ctc-data2vec-ft
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-data2vec-ft
13
+
14
+ This model is a fine-tuned version of [facebook/data2vec-audio-base](https://huggingface.co/facebook/data2vec-audio-base) on an unknown dataset.
15
+ It achieves the following results on the evaluation set:
16
+ - Loss: 3.0464
17
+ - Wer: 1.0
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: 30.0
46
+ - mixed_precision_training: Native AMP
47
+
48
+ ### Training results
49
+
50
+ | Training Loss | Epoch | Step | Validation Loss | Wer |
51
+ |:-------------:|:-----:|:----:|:---------------:|:---:|
52
+ | No log | 0.43 | 100 | 3.3600 | 1.0 |
53
+ | No log | 0.87 | 200 | 3.0887 | 1.0 |
54
+ | No log | 1.3 | 300 | 3.0779 | 1.0 |
55
+ | No log | 1.74 | 400 | 3.0551 | 1.0 |
56
+ | 4.8553 | 2.17 | 500 | 3.0526 | 1.0 |
57
+ | 4.8553 | 2.61 | 600 | 3.0560 | 1.0 |
58
+ | 4.8553 | 3.04 | 700 | 3.1251 | 1.0 |
59
+ | 4.8553 | 3.48 | 800 | 3.0870 | 1.0 |
60
+ | 4.8553 | 3.91 | 900 | 3.0822 | 1.0 |
61
+ | 3.1133 | 4.35 | 1000 | 3.0484 | 1.0 |
62
+ | 3.1133 | 4.78 | 1100 | 3.0558 | 1.0 |
63
+ | 3.1133 | 5.22 | 1200 | 3.1019 | 1.0 |
64
+ | 3.1133 | 5.65 | 1300 | 3.0914 | 1.0 |
65
+ | 3.1133 | 6.09 | 1400 | 3.0691 | 1.0 |
66
+ | 3.109 | 6.52 | 1500 | 3.0589 | 1.0 |
67
+ | 3.109 | 6.95 | 1600 | 3.0508 | 1.0 |
68
+ | 3.109 | 7.39 | 1700 | 3.0540 | 1.0 |
69
+ | 3.109 | 7.82 | 1800 | 3.0546 | 1.0 |
70
+ | 3.109 | 8.26 | 1900 | 3.0524 | 1.0 |
71
+ | 3.1106 | 8.69 | 2000 | 3.0569 | 1.0 |
72
+ | 3.1106 | 9.13 | 2100 | 3.0622 | 1.0 |
73
+ | 3.1106 | 9.56 | 2200 | 3.0518 | 1.0 |
74
+ | 3.1106 | 10.0 | 2300 | 3.0749 | 1.0 |
75
+ | 3.1106 | 10.43 | 2400 | 3.0698 | 1.0 |
76
+ | 3.1058 | 10.87 | 2500 | 3.0665 | 1.0 |
77
+ | 3.1058 | 11.3 | 2600 | 3.0555 | 1.0 |
78
+ | 3.1058 | 11.74 | 2700 | 3.0589 | 1.0 |
79
+ | 3.1058 | 12.17 | 2800 | 3.0611 | 1.0 |
80
+ | 3.1058 | 12.61 | 2900 | 3.0561 | 1.0 |
81
+ | 3.1071 | 13.04 | 3000 | 3.0480 | 1.0 |
82
+ | 3.1071 | 13.48 | 3100 | 3.0492 | 1.0 |
83
+ | 3.1071 | 13.91 | 3200 | 3.0574 | 1.0 |
84
+ | 3.1071 | 14.35 | 3300 | 3.0538 | 1.0 |
85
+ | 3.1071 | 14.78 | 3400 | 3.0505 | 1.0 |
86
+ | 3.1061 | 15.22 | 3500 | 3.0600 | 1.0 |
87
+ | 3.1061 | 15.65 | 3600 | 3.0596 | 1.0 |
88
+ | 3.1061 | 16.09 | 3700 | 3.0623 | 1.0 |
89
+ | 3.1061 | 16.52 | 3800 | 3.0800 | 1.0 |
90
+ | 3.1061 | 16.95 | 3900 | 3.0583 | 1.0 |
91
+ | 3.1036 | 17.39 | 4000 | 3.0534 | 1.0 |
92
+ | 3.1036 | 17.82 | 4100 | 3.0563 | 1.0 |
93
+ | 3.1036 | 18.26 | 4200 | 3.0481 | 1.0 |
94
+ | 3.1036 | 18.69 | 4300 | 3.0477 | 1.0 |
95
+ | 3.1036 | 19.13 | 4400 | 3.0505 | 1.0 |
96
+ | 3.1086 | 19.56 | 4500 | 3.0485 | 1.0 |
97
+ | 3.1086 | 20.0 | 4600 | 3.0481 | 1.0 |
98
+ | 3.1086 | 20.43 | 4700 | 3.0615 | 1.0 |
99
+ | 3.1086 | 20.87 | 4800 | 3.0658 | 1.0 |
100
+ | 3.1086 | 21.3 | 4900 | 3.0505 | 1.0 |
101
+ | 3.1028 | 21.74 | 5000 | 3.0492 | 1.0 |
102
+ | 3.1028 | 22.17 | 5100 | 3.0485 | 1.0 |
103
+ | 3.1028 | 22.61 | 5200 | 3.0483 | 1.0 |
104
+ | 3.1028 | 23.04 | 5300 | 3.0479 | 1.0 |
105
+ | 3.1028 | 23.48 | 5400 | 3.0509 | 1.0 |
106
+ | 3.1087 | 23.91 | 5500 | 3.0530 | 1.0 |
107
+ | 3.1087 | 24.35 | 5600 | 3.0486 | 1.0 |
108
+ | 3.1087 | 24.78 | 5700 | 3.0514 | 1.0 |
109
+ | 3.1087 | 25.22 | 5800 | 3.0505 | 1.0 |
110
+ | 3.1087 | 25.65 | 5900 | 3.0508 | 1.0 |
111
+ | 3.1043 | 26.09 | 6000 | 3.0501 | 1.0 |
112
+ | 3.1043 | 26.52 | 6100 | 3.0467 | 1.0 |
113
+ | 3.1043 | 26.95 | 6200 | 3.0466 | 1.0 |
114
+ | 3.1043 | 27.39 | 6300 | 3.0465 | 1.0 |
115
+ | 3.1043 | 27.82 | 6400 | 3.0465 | 1.0 |
116
+ | 3.1175 | 28.26 | 6500 | 3.0466 | 1.0 |
117
+ | 3.1175 | 28.69 | 6600 | 3.0466 | 1.0 |
118
+ | 3.1175 | 29.13 | 6700 | 3.0465 | 1.0 |
119
+ | 3.1175 | 29.56 | 6800 | 3.0465 | 1.0 |
120
+ | 3.1175 | 30.0 | 6900 | 3.0464 | 1.0 |
121
+
122
+
123
+ ### Framework versions
124
+
125
+ - Transformers 4.18.0.dev0
126
+ - Pytorch 1.10.2+cu113
127
+ - Datasets 1.18.4
128
+ - Tokenizers 0.11.6