Hugging Face's logo

gigant
/
romanian-wav2vec2

Automatic Speech Recognition
Transformers
PyTorch
Safetensors
Romanian
wav2vec2
hf-asr-leaderboard
robust-speech-event
Eval Results
Inference Endpoints
Model card Files Community
2
gigant commited on
Commit
1c912cb
•
1 Parent(s): 1981104

update model card README.md

Browse files
Files changed (1) hide show
  1. README.md +129 -0
README.md ADDED
@@ -0,0 +1,129 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ language:
3
+ - ro
4
+ license: apache-2.0
5
+ tags:
6
+ - automatic-speech-recognition
7
+ - mozilla-foundation/common_voice_8_0
8
+ - generated_from_trainer
9
+ model-index:
10
+ - name: wav2vec2-ro-300m_01
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
+ # wav2vec2-ro-300m_01
18
+
19
+ This model is a fine-tuned version of [facebook/wav2vec2-xls-r-300m](https://huggingface.co/facebook/wav2vec2-xls-r-300m) on the MOZILLA-FOUNDATION/COMMON_VOICE_8_0 - RO dataset.
20
+ It achieves the following results on the evaluation set:
21
+ - Loss: 0.1553
22
+ - Wer: 0.1174
23
+ - Cer: 0.0294
24
+
25
+ ## Model description
26
+
27
+ More information needed
28
+
29
+ ## Intended uses & limitations
30
+
31
+ More information needed
32
+
33
+ ## Training and evaluation data
34
+
35
+ More information needed
36
+
37
+ ## Training procedure
38
+
39
+ ### Training hyperparameters
40
+
41
+ The following hyperparameters were used during training:
42
+ - learning_rate: 0.003
43
+ - train_batch_size: 16
44
+ - eval_batch_size: 8
45
+ - seed: 42
46
+ - gradient_accumulation_steps: 3
47
+ - total_train_batch_size: 48
48
+ - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
49
+ - lr_scheduler_type: linear
50
+ - lr_scheduler_warmup_steps: 500
51
+ - num_epochs: 50.0
52
+ - mixed_precision_training: Native AMP
53
+
54
+ ### Training results
55
+
56
+ | Training Loss | Epoch | Step | Validation Loss | Wer | Cer |
57
+ |:-------------:|:-----:|:-----:|:---------------:|:------:|:------:|
58
+ | 2.9272 | 0.78 | 500 | 0.7603 | 0.7734 | 0.2355 |
59
+ | 0.6157 | 1.55 | 1000 | 0.4003 | 0.4866 | 0.1247 |
60
+ | 0.4452 | 2.33 | 1500 | 0.2960 | 0.3689 | 0.0910 |
61
+ | 0.3631 | 3.11 | 2000 | 0.2580 | 0.3205 | 0.0796 |
62
+ | 0.3153 | 3.88 | 2500 | 0.2465 | 0.2977 | 0.0747 |
63
+ | 0.2795 | 4.66 | 3000 | 0.2274 | 0.2789 | 0.0694 |
64
+ | 0.2615 | 5.43 | 3500 | 0.2277 | 0.2685 | 0.0675 |
65
+ | 0.2389 | 6.21 | 4000 | 0.2135 | 0.2518 | 0.0627 |
66
+ | 0.2229 | 6.99 | 4500 | 0.2054 | 0.2449 | 0.0614 |
67
+ | 0.2067 | 7.76 | 5000 | 0.2096 | 0.2378 | 0.0597 |
68
+ | 0.1977 | 8.54 | 5500 | 0.2042 | 0.2387 | 0.0600 |
69
+ | 0.1896 | 9.32 | 6000 | 0.2110 | 0.2383 | 0.0595 |
70
+ | 0.1801 | 10.09 | 6500 | 0.1909 | 0.2165 | 0.0548 |
71
+ | 0.174 | 10.87 | 7000 | 0.1883 | 0.2206 | 0.0559 |
72
+ | 0.1685 | 11.65 | 7500 | 0.1848 | 0.2097 | 0.0528 |
73
+ | 0.1591 | 12.42 | 8000 | 0.1851 | 0.2039 | 0.0514 |
74
+ | 0.1537 | 13.2 | 8500 | 0.1881 | 0.2065 | 0.0518 |
75
+ | 0.1504 | 13.97 | 9000 | 0.1840 | 0.1972 | 0.0499 |
76
+ | 0.145 | 14.75 | 9500 | 0.1845 | 0.2029 | 0.0517 |
77
+ | 0.1417 | 15.53 | 10000 | 0.1884 | 0.2003 | 0.0507 |
78
+ | 0.1364 | 16.3 | 10500 | 0.2010 | 0.2037 | 0.0517 |
79
+ | 0.1331 | 17.08 | 11000 | 0.1838 | 0.1923 | 0.0483 |
80
+ | 0.129 | 17.86 | 11500 | 0.1818 | 0.1922 | 0.0489 |
81
+ | 0.1198 | 18.63 | 12000 | 0.1760 | 0.1861 | 0.0465 |
82
+ | 0.1203 | 19.41 | 12500 | 0.1686 | 0.1839 | 0.0465 |
83
+ | 0.1225 | 20.19 | 13000 | 0.1828 | 0.1920 | 0.0479 |
84
+ | 0.1145 | 20.96 | 13500 | 0.1673 | 0.1784 | 0.0446 |
85
+ | 0.1053 | 21.74 | 14000 | 0.1802 | 0.1810 | 0.0456 |
86
+ | 0.1071 | 22.51 | 14500 | 0.1769 | 0.1775 | 0.0444 |
87
+ | 0.1053 | 23.29 | 15000 | 0.1920 | 0.1783 | 0.0457 |
88
+ | 0.1024 | 24.07 | 15500 | 0.1904 | 0.1775 | 0.0446 |
89
+ | 0.0987 | 24.84 | 16000 | 0.1793 | 0.1762 | 0.0446 |
90
+ | 0.0949 | 25.62 | 16500 | 0.1801 | 0.1766 | 0.0443 |
91
+ | 0.0942 | 26.4 | 17000 | 0.1731 | 0.1659 | 0.0423 |
92
+ | 0.0906 | 27.17 | 17500 | 0.1776 | 0.1698 | 0.0424 |
93
+ | 0.0861 | 27.95 | 18000 | 0.1716 | 0.1600 | 0.0406 |
94
+ | 0.0851 | 28.73 | 18500 | 0.1662 | 0.1630 | 0.0410 |
95
+ | 0.0844 | 29.5 | 19000 | 0.1671 | 0.1572 | 0.0393 |
96
+ | 0.0792 | 30.28 | 19500 | 0.1768 | 0.1599 | 0.0407 |
97
+ | 0.0798 | 31.06 | 20000 | 0.1732 | 0.1558 | 0.0394 |
98
+ | 0.0779 | 31.83 | 20500 | 0.1694 | 0.1544 | 0.0388 |
99
+ | 0.0718 | 32.61 | 21000 | 0.1709 | 0.1578 | 0.0399 |
100
+ | 0.0732 | 33.38 | 21500 | 0.1697 | 0.1523 | 0.0391 |
101
+ | 0.0708 | 34.16 | 22000 | 0.1616 | 0.1474 | 0.0375 |
102
+ | 0.0678 | 34.94 | 22500 | 0.1698 | 0.1474 | 0.0375 |
103
+ | 0.0642 | 35.71 | 23000 | 0.1681 | 0.1459 | 0.0369 |
104
+ | 0.0661 | 36.49 | 23500 | 0.1612 | 0.1411 | 0.0357 |
105
+ | 0.0629 | 37.27 | 24000 | 0.1662 | 0.1414 | 0.0355 |
106
+ | 0.0587 | 38.04 | 24500 | 0.1659 | 0.1408 | 0.0351 |
107
+ | 0.0581 | 38.82 | 25000 | 0.1612 | 0.1382 | 0.0352 |
108
+ | 0.0556 | 39.6 | 25500 | 0.1647 | 0.1376 | 0.0345 |
109
+ | 0.0543 | 40.37 | 26000 | 0.1658 | 0.1335 | 0.0337 |
110
+ | 0.052 | 41.15 | 26500 | 0.1716 | 0.1369 | 0.0343 |
111
+ | 0.0513 | 41.92 | 27000 | 0.1600 | 0.1317 | 0.0330 |
112
+ | 0.0491 | 42.7 | 27500 | 0.1671 | 0.1311 | 0.0328 |
113
+ | 0.0463 | 43.48 | 28000 | 0.1613 | 0.1289 | 0.0324 |
114
+ | 0.0468 | 44.25 | 28500 | 0.1599 | 0.1260 | 0.0315 |
115
+ | 0.0435 | 45.03 | 29000 | 0.1556 | 0.1232 | 0.0308 |
116
+ | 0.043 | 45.81 | 29500 | 0.1588 | 0.1240 | 0.0309 |
117
+ | 0.0421 | 46.58 | 30000 | 0.1567 | 0.1217 | 0.0308 |
118
+ | 0.04 | 47.36 | 30500 | 0.1533 | 0.1198 | 0.0302 |
119
+ | 0.0389 | 48.14 | 31000 | 0.1582 | 0.1185 | 0.0297 |
120
+ | 0.0387 | 48.91 | 31500 | 0.1576 | 0.1187 | 0.0297 |
121
+ | 0.0376 | 49.69 | 32000 | 0.1560 | 0.1182 | 0.0295 |
122
+
123
+
124
+ ### Framework versions
125
+
126
+ - Transformers 4.16.0.dev0
127
+ - Pytorch 1.10.1+cu102
128
+ - Datasets 1.18.3
129
+ - Tokenizers 0.11.0