Hugging Face's logo

vasilis
/
xls-r-et-V-3

Automatic Speech Recognition
Transformers
PyTorch
TensorBoard
Estonian
wav2vec2
mozilla-foundation/common_voice_8_0
robust-speech-event
Generated from Trainer
hf-asr-leaderboard
Eval Results
Inference Endpoints
Model card Files Metrics Community
vasilis commited on
Commit
c2b6876
•
1 Parent(s): 4b628c5

update model card README.md

Browse files
Files changed (1) hide show
  1. README.md +114 -0
README.md ADDED
@@ -0,0 +1,114 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ language:
3
+ - et
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 - ET dataset.
22
+ It achieves the following results on the evaluation set:
23
+ - Loss: 0.8824
24
+ - Wer: 0.5246
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: 7e-05
44
+ - train_batch_size: 32
45
+ - eval_batch_size: 32
46
+ - seed: 42
47
+ - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
48
+ - lr_scheduler_type: linear
49
+ - lr_scheduler_warmup_steps: 500
50
+ - training_steps: 25000
51
+ - mixed_precision_training: Native AMP
52
+
53
+ ### Training results
54
+
55
+ | Training Loss | Epoch | Step | Validation Loss | Wer |
56
+ |:-------------:|:------:|:-----:|:---------------:|:------:|
57
+ | 1.0296 | 2.79 | 500 | 0.8106 | 0.8029 |
58
+ | 0.9339 | 5.59 | 1000 | 0.7419 | 0.7932 |
59
+ | 0.8925 | 8.38 | 1500 | 0.7137 | 0.7706 |
60
+ | 0.8484 | 11.17 | 2000 | 0.7020 | 0.7677 |
61
+ | 0.7521 | 13.97 | 2500 | 0.7043 | 0.7375 |
62
+ | 0.719 | 16.76 | 3000 | 0.6617 | 0.7428 |
63
+ | 0.656 | 19.55 | 3500 | 0.6388 | 0.7202 |
64
+ | 0.6085 | 22.35 | 4000 | 0.6211 | 0.6960 |
65
+ | 0.5598 | 25.14 | 4500 | 0.6132 | 0.6644 |
66
+ | 0.4969 | 27.93 | 5000 | 0.6065 | 0.6521 |
67
+ | 0.4638 | 30.73 | 5500 | 0.6978 | 0.6577 |
68
+ | 0.4385 | 33.52 | 6000 | 0.5994 | 0.6565 |
69
+ | 0.396 | 36.31 | 6500 | 0.6170 | 0.6258 |
70
+ | 0.3861 | 39.11 | 7000 | 0.6486 | 0.6217 |
71
+ | 0.3602 | 41.9 | 7500 | 0.6508 | 0.6115 |
72
+ | 0.3251 | 44.69 | 8000 | 0.7022 | 0.6253 |
73
+ | 0.3197 | 47.49 | 8500 | 0.7706 | 0.6215 |
74
+ | 0.3013 | 50.28 | 9000 | 0.6419 | 0.5999 |
75
+ | 0.2813 | 53.07 | 9500 | 0.6908 | 0.5959 |
76
+ | 0.286 | 55.87 | 10000 | 0.7151 | 0.5916 |
77
+ | 0.2645 | 58.66 | 10500 | 0.7181 | 0.5860 |
78
+ | 0.2535 | 61.45 | 11000 | 0.7877 | 0.5979 |
79
+ | 0.247 | 64.25 | 11500 | 0.8199 | 0.6129 |
80
+ | 0.2412 | 67.04 | 12000 | 0.7679 | 0.5884 |
81
+ | 0.2404 | 69.83 | 12500 | 0.7266 | 0.5816 |
82
+ | 0.2293 | 72.63 | 13000 | 0.7928 | 0.5795 |
83
+ | 0.2176 | 75.42 | 13500 | 0.7916 | 0.5846 |
84
+ | 0.2143 | 78.21 | 14000 | 0.7954 | 0.5765 |
85
+ | 0.2185 | 81.01 | 14500 | 0.8317 | 0.5907 |
86
+ | 0.2057 | 83.8 | 15000 | 0.8016 | 0.5851 |
87
+ | 0.1895 | 86.59 | 15500 | 0.8080 | 0.5679 |
88
+ | 0.1883 | 89.39 | 16000 | 0.8103 | 0.5712 |
89
+ | 0.1802 | 92.18 | 16500 | 0.8383 | 0.5644 |
90
+ | 0.1826 | 94.97 | 17000 | 0.8799 | 0.5657 |
91
+ | 0.1717 | 97.77 | 17500 | 0.8620 | 0.5709 |
92
+ | 0.1701 | 100.56 | 18000 | 0.8717 | 0.5662 |
93
+ | 0.1623 | 103.35 | 18500 | 0.8534 | 0.5594 |
94
+ | 0.158 | 106.15 | 19000 | 0.8595 | 0.5546 |
95
+ | 0.1508 | 108.94 | 19500 | 0.8574 | 0.5545 |
96
+ | 0.142 | 111.73 | 20000 | 0.8671 | 0.5537 |
97
+ | 0.1395 | 114.53 | 20500 | 0.8436 | 0.5525 |
98
+ | 0.1373 | 117.32 | 21000 | 0.8808 | 0.5482 |
99
+ | 0.1338 | 120.11 | 21500 | 0.9024 | 0.5418 |
100
+ | 0.1278 | 122.91 | 22000 | 0.9143 | 0.5409 |
101
+ | 0.1207 | 125.7 | 22500 | 0.8917 | 0.5358 |
102
+ | 0.1203 | 128.49 | 23000 | 0.9041 | 0.5341 |
103
+ | 0.1083 | 131.28 | 23500 | 0.8884 | 0.5341 |
104
+ | 0.1147 | 134.08 | 24000 | 0.8910 | 0.5255 |
105
+ | 0.1129 | 136.87 | 24500 | 0.8826 | 0.5241 |
106
+ | 0.1029 | 139.66 | 25000 | 0.8824 | 0.5246 |
107
+
108
+
109
+ ### Framework versions
110
+
111
+ - Transformers 4.16.0.dev0
112
+ - Pytorch 1.10.1+cu102
113
+ - Datasets 1.17.1.dev0
114
+ - Tokenizers 0.11.0