Hugging Face's logo

NbAiLab
/
nb-wav2vec2-1b-nynorsk

Automatic Speech Recognition
Transformers
PyTorch
TensorBoard
Norwegian Nynorsk
Norwegian
wav2vec2
NbAiLab/NPSC
nb-NN
Eval Results
Inference Endpoints
Model card Files Metrics Community
versae commited on
Commit
6147168
•
1 Parent(s): bb61a58

Create README.old.md

Browse files
Files changed (1) hide show
  1. README.old.md +133 -0
README.old.md ADDED
@@ -0,0 +1,133 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: apache-2.0
3
+ tags:
4
+ - automatic-speech-recognition
5
+ - NbAiLab/NPSC
6
+ - generated_from_trainer
7
+ model-index:
8
+ - name: XLSR-1B-nynorsk-low
9
+ results:
10
+ - task:
11
+ name: Automatic Speech Recognition
12
+ type: automatic-speech-recognition
13
+ dataset:
14
+ name: NPSC
15
+ type: NbAiLab/NPSC
16
+ args: 16K_mp3_nynorsk
17
+ metrics:
18
+ - name: Test (Nynorsk) WER
19
+ type: wer
20
+ value: 0.11319692134409612
21
+ - name: Test (Nynorsk) CER
22
+ type: cer
23
+ value: 0.040263696587740365
24
+ ---
25
+
26
+ <!-- This model card has been generated automatically according to the information the Trainer had access to. You
27
+ should probably proofread and complete it, then remove this comment. -->
28
+
29
+ # XLSR-1B-nynorsk-low
30
+
31
+ This model is a fine-tuned version of [facebook/wav2vec2-xls-r-1b](https://huggingface.co/facebook/wav2vec2-xls-r-1b) on the NBAILAB/NPSC - 16K_MP3_NYNORSK dataset.
32
+ It achieves the following results on the evaluation set:
33
+ - Loss: 0.2909
34
+ - Wer: 0.1364
35
+
36
+ ## Model description
37
+
38
+ More information needed
39
+
40
+ ## Intended uses & limitations
41
+
42
+ More information needed
43
+
44
+ ## Training and evaluation data
45
+
46
+ More information needed
47
+
48
+ ## Training procedure
49
+
50
+ ### Training hyperparameters
51
+
52
+ The following hyperparameters were used during training:
53
+ - learning_rate: 2e-05
54
+ - train_batch_size: 24
55
+ - eval_batch_size: 24
56
+ - seed: 42
57
+ - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
58
+ - lr_scheduler_type: linear
59
+ - lr_scheduler_warmup_steps: 2000
60
+ - num_epochs: 60.0
61
+ - mixed_precision_training: Native AMP
62
+
63
+ ### Training results
64
+
65
+ | Training Loss | Epoch | Step | Validation Loss | Wer |
66
+ |:-------------:|:-----:|:-----:|:---------------:|:------:|
67
+ | 2.8979 | 1.0 | 500 | 2.9413 | 1.0 |
68
+ | 1.2224 | 2.0 | 1000 | 1.0359 | 0.7802 |
69
+ | 0.8643 | 3.01 | 1500 | 0.7746 | 0.5969 |
70
+ | 0.8211 | 4.01 | 2000 | 0.4882 | 0.3710 |
71
+ | 0.5287 | 5.01 | 2500 | 0.4060 | 0.3085 |
72
+ | 0.4724 | 6.01 | 3000 | 0.3297 | 0.2517 |
73
+ | 0.4357 | 7.01 | 3500 | 0.3106 | 0.2342 |
74
+ | 0.376 | 8.02 | 4000 | 0.2776 | 0.2072 |
75
+ | 0.3286 | 9.02 | 4500 | 0.2888 | 0.2032 |
76
+ | 0.3731 | 10.02 | 5000 | 0.2691 | 0.1835 |
77
+ | 0.306 | 11.02 | 5500 | 0.2536 | 0.1835 |
78
+ | 0.3025 | 12.02 | 6000 | 0.2758 | 0.1809 |
79
+ | 0.3413 | 13.03 | 6500 | 0.2791 | 0.1823 |
80
+ | 0.2601 | 14.03 | 7000 | 0.2912 | 0.1759 |
81
+ | 0.2332 | 15.03 | 7500 | 0.2582 | 0.1694 |
82
+ | 0.2108 | 16.03 | 8000 | 0.2717 | 0.1660 |
83
+ | 0.2122 | 17.03 | 8500 | 0.2848 | 0.1647 |
84
+ | 0.2369 | 18.04 | 9000 | 0.2548 | 0.1646 |
85
+ | 0.1906 | 19.04 | 9500 | 0.2667 | 0.1627 |
86
+ | 0.1943 | 20.04 | 10000 | 0.2662 | 0.1623 |
87
+ | 0.18 | 21.04 | 10500 | 0.2769 | 0.1561 |
88
+ | 0.1654 | 22.04 | 11000 | 0.2661 | 0.1558 |
89
+ | 0.1515 | 23.05 | 11500 | 0.2870 | 0.1597 |
90
+ | 0.147 | 24.05 | 12000 | 0.2778 | 0.1551 |
91
+ | 0.1622 | 25.05 | 12500 | 0.2753 | 0.1541 |
92
+ | 0.1522 | 26.05 | 13000 | 0.2932 | 0.1521 |
93
+ | 0.1522 | 27.05 | 13500 | 0.2548 | 0.1513 |
94
+ | 0.1319 | 28.06 | 14000 | 0.2811 | 0.1532 |
95
+ | 0.1261 | 29.06 | 14500 | 0.2786 | 0.1521 |
96
+ | 0.1391 | 30.06 | 15000 | 0.2651 | 0.1461 |
97
+ | 0.1486 | 31.06 | 15500 | 0.2866 | 0.1494 |
98
+ | 0.1121 | 32.06 | 16000 | 0.2641 | 0.1478 |
99
+ | 0.1114 | 33.07 | 16500 | 0.2910 | 0.1478 |
100
+ | 0.101 | 34.07 | 17000 | 0.2884 | 0.1443 |
101
+ | 0.1135 | 35.07 | 17500 | 0.3029 | 0.1469 |
102
+ | 0.0972 | 36.07 | 18000 | 0.2870 | 0.1467 |
103
+ | 0.1178 | 37.07 | 18500 | 0.2745 | 0.1450 |
104
+ | 0.0885 | 38.08 | 19000 | 0.2836 | 0.1440 |
105
+ | 0.1144 | 39.08 | 19500 | 0.2761 | 0.1446 |
106
+ | 0.0997 | 40.08 | 20000 | 0.2806 | 0.1439 |
107
+ | 0.1012 | 41.08 | 20500 | 0.2878 | 0.1413 |
108
+ | 0.0902 | 42.08 | 21000 | 0.2832 | 0.1452 |
109
+ | 0.0804 | 43.09 | 21500 | 0.2911 | 0.1458 |
110
+ | 0.0762 | 44.09 | 22000 | 0.2708 | 0.1441 |
111
+ | 0.0758 | 45.09 | 22500 | 0.2804 | 0.1434 |
112
+ | 0.0874 | 46.09 | 23000 | 0.2831 | 0.1407 |
113
+ | 0.0895 | 47.09 | 23500 | 0.2913 | 0.1396 |
114
+ | 0.0975 | 48.1 | 24000 | 0.2956 | 0.1411 |
115
+ | 0.0758 | 49.1 | 24500 | 0.2920 | 0.1385 |
116
+ | 0.0704 | 50.1 | 25000 | 0.2788 | 0.1383 |
117
+ | 0.0707 | 51.1 | 25500 | 0.2822 | 0.1388 |
118
+ | 0.0664 | 52.1 | 26000 | 0.2876 | 0.1371 |
119
+ | 0.0692 | 53.11 | 26500 | 0.2815 | 0.1377 |
120
+ | 0.0799 | 54.11 | 27000 | 0.2806 | 0.1363 |
121
+ | 0.0611 | 55.11 | 27500 | 0.2878 | 0.1363 |
122
+ | 0.0759 | 56.11 | 28000 | 0.2900 | 0.1365 |
123
+ | 0.0801 | 57.11 | 28500 | 0.2881 | 0.1375 |
124
+ | 0.0644 | 58.12 | 29000 | 0.2898 | 0.1362 |
125
+ | 0.068 | 59.12 | 29500 | 0.2913 | 0.1369 |
126
+
127
+
128
+ ### Framework versions
129
+
130
+ - Transformers 4.17.0.dev0
131
+ - Pytorch 1.10.2+cu113
132
+ - Datasets 2.0.0
133
+ - Tokenizers 0.11.0