RuudVelo commited on
Commit
40e618b
1 Parent(s): e729fb8

Upload README.md

Browse files
Files changed (1) hide show
  1. README.md +162 -0
README.md ADDED
@@ -0,0 +1,162 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ language:
3
+ - nl
4
+ license: apache-2.0
5
+ tags:
6
+ - automatic-speech-recognition
7
+ - mozilla-foundation/common_voice_8_0
8
+ - generated_from_trainer
9
+ datasets:
10
+ - common_voice
11
+ model-index:
12
+ - name: ''
13
+ results: []
14
+ ---
15
+
16
+ <!-- This model card has been generated automatically according to the information the Trainer had access to. You
17
+ should probably proofread and complete it, then remove this comment. -->
18
+
19
+ #
20
+
21
+ This model is a fine-tuned version of [facebook/wav2vec2-xls-r-1b](https://huggingface.co/facebook/wav2vec2-xls-r-1b) on the MOZILLA-FOUNDATION/COMMON_VOICE_8_0 - NL dataset.
22
+ It achieves the following results on the evaluation set:
23
+ - Loss: 0.1479
24
+ - Wer: 0.1156
25
+
26
+ ## Model description
27
+
28
+ More information needed
29
+
30
+ ## Intended uses & limitations
31
+
32
+ More information needed
33
+
34
+ ## Training and evaluation data
35
+
36
+ More information needed
37
+
38
+ ## Training procedure
39
+
40
+ ### Training hyperparameters
41
+
42
+ The following hyperparameters were used during training:
43
+ - learning_rate: 5e-05
44
+ - train_batch_size: 8
45
+ - eval_batch_size: 8
46
+ - seed: 42
47
+ - gradient_accumulation_steps: 5
48
+ - total_train_batch_size: 40
49
+ - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
50
+ - lr_scheduler_type: linear
51
+ - lr_scheduler_warmup_steps: 500
52
+ - num_epochs: 50.0
53
+ - mixed_precision_training: Native AMP
54
+
55
+ ### Training results
56
+
57
+ | Training Loss | Epoch | Step | Validation Loss | Wer |
58
+ |:-------------:|:-----:|:-----:|:---------------:|:------:|
59
+ | 1.2223 | 0.52 | 500 | 0.3866 | 0.3425 |
60
+ | 1.0748 | 1.03 | 1000 | 0.2574 | 0.2169 |
61
+ | 1.0416 | 1.55 | 1500 | 0.2177 | 0.1946 |
62
+ | 0.9951 | 2.06 | 2000 | 0.2008 | 0.1760 |
63
+ | 0.975 | 2.58 | 2500 | 0.1961 | 0.1751 |
64
+ | 0.9461 | 3.1 | 3000 | 0.1989 | 0.1782 |
65
+ | 0.9381 | 3.61 | 3500 | 0.1928 | 0.1699 |
66
+ | 0.934 | 4.13 | 4000 | 0.1923 | 0.1633 |
67
+ | 0.9322 | 4.64 | 4500 | 0.1871 | 0.1634 |
68
+ | 0.9012 | 5.16 | 5000 | 0.1890 | 0.1702 |
69
+ | 0.9045 | 5.68 | 5500 | 0.1882 | 0.1740 |
70
+ | 0.8826 | 6.19 | 6000 | 0.1856 | 0.1575 |
71
+ | 0.8848 | 6.71 | 6500 | 0.1861 | 0.1617 |
72
+ | 0.8723 | 7.22 | 7000 | 0.1927 | 0.1646 |
73
+ | 0.8725 | 7.74 | 7500 | 0.1798 | 0.1531 |
74
+ | 0.8573 | 8.26 | 8000 | 0.1781 | 0.1587 |
75
+ | 0.8633 | 8.77 | 8500 | 0.1852 | 0.1628 |
76
+ | 0.8603 | 9.29 | 9000 | 0.1833 | 0.1601 |
77
+ | 0.8421 | 9.8 | 9500 | 0.1788 | 0.1543 |
78
+ | 0.8404 | 10.32 | 10000 | 0.1844 | 0.1556 |
79
+ | 0.8342 | 10.84 | 10500 | 0.1770 | 0.1538 |
80
+ | 0.8161 | 11.35 | 11000 | 0.1821 | 0.1567 |
81
+ | 0.8371 | 11.87 | 11500 | 0.1909 | 0.1629 |
82
+ | 0.8083 | 12.38 | 12000 | 0.1778 | 0.1498 |
83
+ | 0.806 | 12.9 | 12500 | 0.1802 | 0.1547 |
84
+ | 0.8013 | 13.42 | 13000 | 0.1859 | 0.1584 |
85
+ | 0.7913 | 13.93 | 13500 | 0.1875 | 0.1517 |
86
+ | 0.8063 | 14.45 | 14000 | 0.1799 | 0.1571 |
87
+ | 0.7991 | 14.96 | 14500 | 0.1792 | 0.1538 |
88
+ | 0.7843 | 15.48 | 15000 | 0.1753 | 0.1464 |
89
+ | 0.7905 | 16.0 | 15500 | 0.1784 | 0.1508 |
90
+ | 0.7808 | 16.51 | 16000 | 0.1771 | 0.1485 |
91
+ | 0.7743 | 17.03 | 16500 | 0.1795 | 0.1491 |
92
+ | 0.7833 | 17.54 | 17000 | 0.1722 | 0.1484 |
93
+ | 0.7763 | 18.06 | 17500 | 0.1767 | 0.1518 |
94
+ | 0.7698 | 18.58 | 18000 | 0.1720 | 0.1460 |
95
+ | 0.7571 | 19.09 | 18500 | 0.1735 | 0.1478 |
96
+ | 0.7673 | 19.61 | 19000 | 0.1817 | 0.1511 |
97
+ | 0.7415 | 20.12 | 19500 | 0.1763 | 0.1481 |
98
+ | 0.751 | 20.64 | 20000 | 0.1742 | 0.1484 |
99
+ | 0.7563 | 21.16 | 20500 | 0.1810 | 0.1611 |
100
+ | 0.7423 | 21.67 | 21000 | 0.1817 | 0.1557 |
101
+ | 0.7242 | 22.19 | 21500 | 0.1690 | 0.1446 |
102
+ | 0.7251 | 22.7 | 22000 | 0.1684 | 0.1446 |
103
+ | 0.7302 | 23.22 | 22500 | 0.1735 | 0.1430 |
104
+ | 0.733 | 23.74 | 23000 | 0.1720 | 0.1454 |
105
+ | 0.7128 | 24.25 | 23500 | 0.1668 | 0.1383 |
106
+ | 0.7184 | 24.77 | 24000 | 0.1635 | 0.1377 |
107
+ | 0.7015 | 25.28 | 24500 | 0.1646 | 0.1389 |
108
+ | 0.7198 | 25.8 | 25000 | 0.1775 | 0.1462 |
109
+ | 0.7178 | 26.32 | 25500 | 0.1705 | 0.1419 |
110
+ | 0.7199 | 26.83 | 26000 | 0.1649 | 0.1416 |
111
+ | 0.6981 | 27.35 | 26500 | 0.1724 | 0.1418 |
112
+ | 0.6886 | 27.86 | 27000 | 0.1633 | 0.1382 |
113
+ | 0.6922 | 28.38 | 27500 | 0.1698 | 0.1420 |
114
+ | 0.6833 | 28.9 | 28000 | 0.1611 | 0.1351 |
115
+ | 0.6798 | 29.41 | 28500 | 0.1639 | 0.1365 |
116
+ | 0.6711 | 29.93 | 29000 | 0.1668 | 0.1358 |
117
+ | 0.6762 | 30.44 | 29500 | 0.1682 | 0.1355 |
118
+ | 0.6594 | 30.96 | 30000 | 0.1629 | 0.1345 |
119
+ | 0.6664 | 31.48 | 30500 | 0.1625 | 0.1321 |
120
+ | 0.6838 | 31.99 | 31000 | 0.1597 | 0.1372 |
121
+ | 0.6603 | 32.51 | 31500 | 0.1583 | 0.1302 |
122
+ | 0.6468 | 33.02 | 32000 | 0.1595 | 0.1322 |
123
+ | 0.6464 | 33.54 | 32500 | 0.1609 | 0.1315 |
124
+ | 0.6623 | 34.06 | 33000 | 0.1622 | 0.1366 |
125
+ | 0.6414 | 34.57 | 33500 | 0.1587 | 0.1330 |
126
+ | 0.6242 | 35.09 | 34000 | 0.1614 | 0.1337 |
127
+ | 0.632 | 35.6 | 34500 | 0.1568 | 0.1272 |
128
+ | 0.6346 | 36.12 | 35000 | 0.1583 | 0.1274 |
129
+ | 0.6143 | 36.64 | 35500 | 0.1576 | 0.1264 |
130
+ | 0.6208 | 37.15 | 36000 | 0.1621 | 0.1263 |
131
+ | 0.6185 | 37.67 | 36500 | 0.1623 | 0.1270 |
132
+ | 0.6128 | 38.18 | 37000 | 0.1604 | 0.1268 |
133
+ | 0.6151 | 38.7 | 37500 | 0.1593 | 0.1246 |
134
+ | 0.6082 | 39.22 | 38000 | 0.1532 | 0.1238 |
135
+ | 0.6 | 39.73 | 38500 | 0.1524 | 0.1224 |
136
+ | 0.6032 | 40.25 | 39000 | 0.1521 | 0.1212 |
137
+ | 0.6016 | 40.76 | 39500 | 0.1551 | 0.1215 |
138
+ | 0.6009 | 41.28 | 40000 | 0.1523 | 0.1215 |
139
+ | 0.5875 | 41.8 | 40500 | 0.1541 | 0.1216 |
140
+ | 0.608 | 42.31 | 41000 | 0.1536 | 0.1209 |
141
+ | 0.5876 | 42.83 | 41500 | 0.1567 | 0.1211 |
142
+ | 0.5714 | 43.34 | 42000 | 0.1532 | 0.1217 |
143
+ | 0.5756 | 43.86 | 42500 | 0.1516 | 0.1196 |
144
+ | 0.5719 | 44.38 | 43000 | 0.1491 | 0.1191 |
145
+ | 0.5829 | 44.89 | 43500 | 0.1497 | 0.1193 |
146
+ | 0.5664 | 45.41 | 44000 | 0.1487 | 0.1173 |
147
+ | 0.5707 | 45.92 | 44500 | 0.1470 | 0.1164 |
148
+ | 0.5696 | 46.44 | 45000 | 0.1479 | 0.1161 |
149
+ | 0.5767 | 46.96 | 45500 | 0.1492 | 0.1175 |
150
+ | 0.5573 | 47.47 | 46000 | 0.1471 | 0.1165 |
151
+ | 0.5625 | 47.99 | 46500 | 0.1484 | 0.1168 |
152
+ | 0.5671 | 48.5 | 47000 | 0.1474 | 0.1162 |
153
+ | 0.5484 | 49.02 | 47500 | 0.1479 | 0.1158 |
154
+ | 0.555 | 49.54 | 48000 | 0.1477 | 0.1157 |
155
+
156
+
157
+ ### Framework versions
158
+
159
+ - Transformers 4.17.0.dev0
160
+ - Pytorch 1.10.2+cu102
161
+ - Datasets 1.18.2.dev0
162
+ - Tokenizers 0.11.0