File size: 12,705 Bytes
9c12b5f
9f4c134
 
9c12b5f
 
9f4c134
 
9c12b5f
9f4c134
 
 
9c12b5f
d8faf6a
9c12b5f
44c7a85
9f4c134
 
 
 
 
d8faf6a
9f4c134
 
 
 
 
95df7a0
9f4c134
 
95df7a0
9f4c134
 
 
 
 
 
 
 
 
 
95df7a0
9f4c134
 
95df7a0
9c12b5f
 
 
 
 
44c7a85
9c12b5f
 
44c7a85
d8faf6a
 
9c12b5f
 
 
e07dce4
9c12b5f
 
 
 
 
 
 
e07dce4
9c12b5f
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
---
language:
- nl
license: apache-2.0
tags:
- automatic-speech-recognition
- mozilla-foundation/common_voice_8_0
- generated_from_trainer
- nl
- robust-speech-event
- model_for_talk
datasets:
- mozilla-foundation/common_voice_8_0
model-index:
- name: wav2vec2-large-xls-r-300m-cv8-nl
  results:
  - task: 
      name: Automatic Speech Recognition 
      type: automatic-speech-recognition
    dataset:
      name: Common Voice 8
      type: mozilla-foundation/common_voice_8_0
      args: nl
    metrics:
       - name: Test WER
         type: wer
         value: 17.56
       - name: Test CER
         type: cer
         value: 5.49
  - task: 
      name: Automatic Speech Recognition
      type: automatic-speech-recognition
    dataset:
      name: Robust Speech Event - Dev Data
      type: speech-recognition-community-v2/dev_data
      args: nl
    metrics:
       - name: Test WER
         type: wer
         value: 39.25
       - name: Test CER
         type: cer
         value: 16.64
---

<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->

# wav2vec2-large-xls-r-300m-cv8-nl

This model is a fine-tuned version of [facebook/wav2vec2-xls-r-300m](https://huggingface.co/facebook/wav2vec2-xls-r-300m) on the common_voice dataset.
It achieves the following results on the testset of commonvoice:
- Loss: NA, see below loss on validation (eval)
- Wer: 0.371 -> detailed metrics

## Model description

Dutch wav2vec2-xls-r-300m model

## Intended uses & limitations

More information needed

## Training and evaluation data

The model was trained on Dutch common voice 8 with 75 epochs. The train set consisted of the common voice 8 train set and evaluation set was the common voice 8 validation set. The WER reported is on the common voice 8 test set which was not part of training nor validation (eval)

## Training procedure

### Training hyperparameters

The following hyperparameters were used during training:
- learning_rate: 7.5e-05
- train_batch_size: 16
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 32
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 500
- num_epochs: 75
- mixed_precision_training: Native AMP

### Training results

| Training Loss | Epoch | Step  | Validation Loss | Wer    |
|:-------------:|:-----:|:-----:|:---------------:|:------:|
| 4.4631        | 0.45  | 400   | 3.0212          | 1.0    |
| 2.8895        | 0.9   | 800   | 2.8555          | 1.0    |
| 2.5142        | 1.36  | 1200  | 0.8149          | 0.6375 |
| 1.5593        | 1.81  | 1600  | 0.3854          | 0.3574 |
| 1.3414        | 2.26  | 2000  | 0.3059          | 0.3060 |
| 1.2564        | 2.71  | 2400  | 0.2728          | 0.2898 |
| 1.2059        | 3.17  | 2800  | 0.2637          | 0.2748 |
| 1.1632        | 3.62  | 3200  | 0.2366          | 0.2637 |
| 1.1177        | 4.07  | 3600  | 0.2285          | 0.2478 |
| 1.097         | 4.52  | 4000  | 0.2194          | 0.2408 |
| 1.086         | 4.98  | 4400  | 0.2138          | 0.2340 |
| 1.0584        | 5.43  | 4800  | 0.2100          | 0.2294 |
| 1.0539        | 5.88  | 5200  | 0.2033          | 0.2285 |
| 1.042         | 6.33  | 5600  | 0.2066          | 0.2320 |
| 1.0253        | 6.79  | 6000  | 0.2012          | 0.2260 |
| 1.0115        | 7.24  | 6400  | 0.1960          | 0.2201 |
| 1.007         | 7.69  | 6800  | 0.1983          | 0.2210 |
| 0.9987        | 8.14  | 7200  | 0.1955          | 0.2212 |
| 0.9864        | 8.6   | 7600  | 0.1879          | 0.2157 |
| 0.9831        | 9.05  | 8000  | 0.1910          | 0.2221 |
| 0.9707        | 9.5   | 8400  | 0.1944          | 0.2199 |
| 0.9717        | 9.95  | 8800  | 0.1885          | 0.2133 |
| 0.9536        | 10.41 | 9200  | 0.1981          | 0.2157 |
| 0.9551        | 10.86 | 9600  | 0.1880          | 0.2184 |
| 0.9449        | 11.31 | 10000 | 0.1938          | 0.2233 |
| 0.9393        | 11.76 | 10400 | 0.1926          | 0.2136 |
| 0.9348        | 12.22 | 10800 | 0.1893          | 0.2143 |
| 0.9379        | 12.67 | 11200 | 0.1935          | 0.2206 |
| 0.931         | 13.12 | 11600 | 0.1868          | 0.2148 |
| 0.9171        | 13.57 | 12000 | 0.1880          | 0.2085 |
| 0.92          | 14.03 | 12400 | 0.1963          | 0.2132 |
| 0.9057        | 14.48 | 12800 | 0.1881          | 0.2131 |
| 0.9074        | 14.93 | 13200 | 0.1839          | 0.2146 |
| 0.8979        | 15.38 | 13600 | 0.1834          | 0.1954 |
| 0.9079        | 15.84 | 14000 | 0.1800          | 0.1899 |
| 0.8863        | 16.29 | 14400 | 0.1845          | 0.2058 |
| 0.8931        | 16.74 | 14800 | 0.1902          | 0.1973 |
| 0.8859        | 17.19 | 15200 | 0.1956          | 0.1982 |
| 0.8898        | 17.65 | 15600 | 0.1784          | 0.1946 |
| 0.8827        | 18.1  | 16000 | 0.1883          | 0.2067 |
| 0.8882        | 18.55 | 16400 | 0.1851          | 0.2006 |
| 0.884         | 19.0  | 16800 | 0.1877          | 0.1978 |
| 0.8717        | 19.46 | 17200 | 0.1804          | 0.1994 |
| 0.8642        | 19.91 | 17600 | 0.1758          | 0.1987 |
| 0.8635        | 20.36 | 18000 | 0.1840          | 0.2003 |
| 0.8687        | 20.81 | 18400 | 0.1782          | 0.2082 |
| 0.8674        | 21.27 | 18800 | 0.1803          | 0.2046 |
| 0.8555        | 21.72 | 19200 | 0.1858          | 0.2059 |
| 0.8542        | 22.17 | 19600 | 0.1850          | 0.1958 |
| 0.8551        | 22.62 | 20000 | 0.1825          | 0.1946 |
| 0.8424        | 23.08 | 20400 | 0.1827          | 0.1726 |
| 0.8424        | 23.53 | 20800 | 0.1843          | 0.1936 |
| 0.8498        | 23.98 | 21200 | 0.1810          | 0.1985 |
| 0.8299        | 24.43 | 21600 | 0.1774          | 0.1888 |
| 0.8361        | 24.89 | 22000 | 0.1927          | 0.1942 |
| 0.841         | 25.34 | 22400 | 0.1871          | 0.1903 |
| 0.8277        | 25.79 | 22800 | 0.1786          | 0.1867 |
| 0.8272        | 26.24 | 23200 | 0.1893          | 0.1616 |
| 0.8321        | 26.7  | 23600 | 0.1856          | 0.1521 |
| 0.8321        | 27.15 | 24000 | 0.1807          | 0.1477 |
| 0.8212        | 27.6  | 24400 | 0.1777          | 0.1508 |
| 0.8238        | 28.05 | 24800 | 0.1829          | 0.1539 |
| 0.8158        | 28.51 | 25200 | 0.1888          | 0.1619 |
| 0.8042        | 28.96 | 25600 | 0.1864          | 0.1510 |
| 0.8141        | 29.41 | 26000 | 0.1909          | 0.1548 |
| 0.8119        | 29.86 | 26400 | 0.1842          | 0.1523 |
| 0.8023        | 30.32 | 26800 | 0.1852          | 0.1459 |
| 0.8043        | 30.77 | 27200 | 0.1747          | 0.1496 |
| 0.8082        | 31.22 | 27600 | 0.1827          | 0.1512 |
| 0.8011        | 31.67 | 28000 | 0.1850          | 0.1480 |
| 0.7869        | 32.13 | 28400 | 0.1816          | 0.1502 |
| 0.7975        | 32.58 | 28800 | 0.1832          | 0.1511 |
| 0.7811        | 33.03 | 29200 | 0.1810          | 0.1429 |
| 0.7982        | 33.48 | 29600 | 0.1706          | 0.1407 |
| 0.8007        | 33.94 | 30000 | 0.1844          | 0.1548 |
| 0.7907        | 34.39 | 30400 | 0.1843          | 0.1539 |
| 0.8005        | 34.84 | 30800 | 0.1798          | 0.1462 |
| 0.7769        | 35.29 | 31200 | 0.1798          | 0.1494 |
| 0.7869        | 35.75 | 31600 | 0.1868          | 0.1643 |
| 0.7789        | 36.2  | 32000 | 0.1817          | 0.1477 |
| 0.7881        | 36.65 | 32400 | 0.1801          | 0.1419 |
| 0.7832        | 37.1  | 32800 | 0.1765          | 0.1454 |
| 0.778         | 37.56 | 33200 | 0.1779          | 0.1467 |
| 0.779         | 38.01 | 33600 | 0.1829          | 0.1565 |
| 0.7693        | 38.46 | 34000 | 0.1748          | 0.1583 |
| 0.7765        | 38.91 | 34400 | 0.1842          | 0.1683 |
| 0.7786        | 39.37 | 34800 | 0.1897          | 0.1543 |
| 0.7652        | 39.82 | 35200 | 0.1861          | 0.1495 |
| 0.773         | 40.27 | 35600 | 0.1775          | 0.1419 |
| 0.7625        | 40.72 | 36000 | 0.1916          | 0.1525 |
| 0.7625        | 41.18 | 36400 | 0.1800          | 0.1429 |
| 0.7548        | 41.63 | 36800 | 0.1788          | 0.1464 |
| 0.7608        | 42.08 | 37200 | 0.1841          | 0.1457 |
| 0.7614        | 42.53 | 37600 | 0.1805          | 0.1401 |
| 0.7646        | 42.99 | 38000 | 0.1863          | 0.1455 |
| 0.7488        | 43.44 | 38400 | 0.1903          | 0.1479 |
| 0.7566        | 43.89 | 38800 | 0.1825          | 0.1414 |
| 0.7495        | 44.34 | 39200 | 0.1873          | 0.1476 |
| 0.7453        | 44.8  | 39600 | 0.1887          | 0.1473 |
| 0.7414        | 45.25 | 40000 | 0.1783          | 0.1430 |
| 0.7431        | 45.7  | 40400 | 0.1866          | 0.1459 |
| 0.7405        | 46.15 | 40800 | 0.1847          | 0.1442 |
| 0.7421        | 46.61 | 41200 | 0.1824          | 0.1626 |
| 0.7423        | 47.06 | 41600 | 0.1843          | 0.1443 |
| 0.7405        | 47.51 | 42000 | 0.1787          | 0.1444 |
| 0.7339        | 47.96 | 42400 | 0.1764          | 0.1646 |
| 0.7297        | 48.42 | 42800 | 0.1749          | 0.1430 |
| 0.7397        | 48.87 | 43200 | 0.1823          | 0.1518 |
| 0.7328        | 49.32 | 43600 | 0.1838          | 0.1565 |
| 0.7342        | 49.77 | 44000 | 0.1797          | 0.1628 |
| 0.7408        | 50.23 | 44400 | 0.1771          | 0.1641 |
| 0.7286        | 50.68 | 44800 | 0.1826          | 0.1692 |
| 0.7305        | 51.13 | 45200 | 0.1760          | 0.1673 |
| 0.721         | 51.58 | 45600 | 0.1769          | 0.1611 |
| 0.7354        | 52.04 | 46000 | 0.1836          | 0.1604 |
| 0.7181        | 52.49 | 46400 | 0.1777          | 0.1576 |
| 0.7212        | 52.94 | 46800 | 0.1809          | 0.1461 |
| 0.7177        | 53.39 | 47200 | 0.1768          | 0.1430 |
| 0.7173        | 53.85 | 47600 | 0.1759          | 0.1388 |
| 0.7135        | 54.3  | 48000 | 0.1668          | 0.1325 |
| 0.7114        | 54.75 | 48400 | 0.1793          | 0.1422 |
| 0.7104        | 55.2  | 48800 | 0.1735          | 0.1440 |
| 0.7135        | 55.66 | 49200 | 0.1795          | 0.1577 |
| 0.7096        | 56.11 | 49600 | 0.1803          | 0.1589 |
| 0.709         | 56.56 | 50000 | 0.1790          | 0.1651 |
| 0.7044        | 57.01 | 50400 | 0.1795          | 0.1632 |
| 0.7081        | 57.47 | 50800 | 0.1738          | 0.1490 |
| 0.6993        | 57.92 | 51200 | 0.1745          | 0.1406 |
| 0.6972        | 58.37 | 51600 | 0.1734          | 0.1380 |
| 0.6984        | 58.82 | 52000 | 0.1799          | 0.1402 |
| 0.7066        | 59.28 | 52400 | 0.1727          | 0.1381 |
| 0.7046        | 59.73 | 52800 | 0.1760          | 0.1360 |
| 0.7024        | 60.18 | 53200 | 0.1793          | 0.1526 |
| 0.6951        | 60.63 | 53600 | 0.1832          | 0.1598 |
| 0.6987        | 61.09 | 54000 | 0.1771          | 0.1563 |
| 0.6966        | 61.54 | 54400 | 0.1768          | 0.1388 |
| 0.6937        | 61.99 | 54800 | 0.1728          | 0.1374 |
| 0.6882        | 62.44 | 55200 | 0.1782          | 0.1385 |
| 0.6919        | 62.9  | 55600 | 0.1781          | 0.1395 |
| 0.6856        | 63.35 | 56000 | 0.1721          | 0.1351 |
| 0.6948        | 63.8  | 56400 | 0.1761          | 0.1383 |
| 0.6947        | 64.25 | 56800 | 0.1701          | 0.1352 |
| 0.6831        | 64.71 | 57200 | 0.1751          | 0.1371 |
| 0.6858        | 65.16 | 57600 | 0.1704          | 0.1383 |
| 0.6787        | 65.61 | 58000 | 0.1730          | 0.1457 |
| 0.6897        | 66.06 | 58400 | 0.1728          | 0.1412 |
| 0.6845        | 66.52 | 58800 | 0.1734          | 0.1394 |
| 0.6763        | 66.97 | 59200 | 0.1741          | 0.1408 |
| 0.6801        | 67.42 | 59600 | 0.1742          | 0.1460 |
| 0.6901        | 67.87 | 60000 | 0.1755          | 0.1449 |
| 0.6802        | 68.33 | 60400 | 0.1743          | 0.1424 |
| 0.6791        | 68.78 | 60800 | 0.1721          | 0.1359 |
| 0.6819        | 69.23 | 61200 | 0.1749          | 0.1363 |
| 0.6794        | 69.68 | 61600 | 0.1770          | 0.1369 |
| 0.6734        | 70.14 | 62000 | 0.1756          | 0.1353 |
| 0.6811        | 70.59 | 62400 | 0.1777          | 0.1371 |
| 0.6813        | 71.04 | 62800 | 0.1763          | 0.1362 |
| 0.6675        | 71.49 | 63200 | 0.1769          | 0.1372 |
| 0.668         | 71.95 | 63600 | 0.1751          | 0.1368 |
| 0.6695        | 72.4  | 64000 | 0.1757          | 0.1370 |
| 0.668         | 72.85 | 64400 | 0.1758          | 0.1363 |
| 0.667         | 73.3  | 64800 | 0.1769          | 0.1363 |
| 0.6634        | 73.76 | 65200 | 0.1763          | 0.1361 |
| 0.676         | 74.21 | 65600 | 0.1751          | 0.1358 |
| 0.667         | 74.66 | 66000 | 0.1755          | 0.1362 |


### Framework versions

- Transformers 4.16.0.dev0
- Pytorch 1.10.1+cu102
- Datasets 1.17.1.dev0
- Tokenizers 0.11.0