Hugging Face's logo

DrishtiSharma
/
wav2vec2-large-xls-r-300m-br-d2

Automatic Speech Recognition
Transformers
PyTorch
TensorBoard
Breton
wav2vec2
Generated from Trainer
robust-speech-event
hf-asr-leaderboard
Eval Results
Inference Endpoints
Model card Files Metrics Community
DrishtiSharma commited on
Commit
246d6b7
•
1 Parent(s): 2509b61

update model card README.md

Browse files
Files changed (1) hide show
  1. README.md +136 -0
README.md ADDED
@@ -0,0 +1,136 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: apache-2.0
3
+ tags:
4
+ - generated_from_trainer
5
+ datasets:
6
+ - common_voice
7
+ model-index:
8
+ - name: wav2vec2-large-xls-r-300m-br-d2
9
+ results: []
10
+ ---
11
+
12
+ <!-- This model card has been generated automatically according to the information the Trainer had access to. You
13
+ should probably proofread and complete it, then remove this comment. -->
14
+
15
+ # wav2vec2-large-xls-r-300m-br-d2
16
+
17
+ 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.
18
+ It achieves the following results on the evaluation set:
19
+ - Loss: 1.1257
20
+ - Wer: 0.4631
21
+
22
+ ## Model description
23
+
24
+ More information needed
25
+
26
+ ## Intended uses & limitations
27
+
28
+ More information needed
29
+
30
+ ## Training and evaluation data
31
+
32
+ More information needed
33
+
34
+ ## Training procedure
35
+
36
+ ### Training hyperparameters
37
+
38
+ The following hyperparameters were used during training:
39
+ - learning_rate: 0.00034
40
+ - train_batch_size: 16
41
+ - eval_batch_size: 8
42
+ - seed: 42
43
+ - gradient_accumulation_steps: 2
44
+ - total_train_batch_size: 32
45
+ - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
46
+ - lr_scheduler_type: linear
47
+ - lr_scheduler_warmup_steps: 750
48
+ - num_epochs: 50
49
+ - mixed_precision_training: Native AMP
50
+
51
+ ### Training results
52
+
53
+ | Training Loss | Epoch | Step | Validation Loss | Wer |
54
+ |:-------------:|:-----:|:----:|:---------------:|:------:|
55
+ | 14.0379 | 0.68 | 100 | 5.6808 | 1.0 |
56
+ | 3.9145 | 1.35 | 200 | 3.1970 | 1.0 |
57
+ | 3.0293 | 2.03 | 300 | 2.9513 | 1.0 |
58
+ | 2.0927 | 2.7 | 400 | 1.4545 | 0.8887 |
59
+ | 1.1556 | 3.38 | 500 | 1.0966 | 0.7564 |
60
+ | 0.9628 | 4.05 | 600 | 0.9808 | 0.7364 |
61
+ | 0.7869 | 4.73 | 700 | 1.0488 | 0.7355 |
62
+ | 0.703 | 5.41 | 800 | 0.9500 | 0.6881 |
63
+ | 0.6657 | 6.08 | 900 | 0.9309 | 0.6259 |
64
+ | 0.5663 | 6.76 | 1000 | 0.9133 | 0.6357 |
65
+ | 0.496 | 7.43 | 1100 | 0.9890 | 0.6028 |
66
+ | 0.4748 | 8.11 | 1200 | 0.9469 | 0.5894 |
67
+ | 0.4135 | 8.78 | 1300 | 0.9270 | 0.6045 |
68
+ | 0.3579 | 9.46 | 1400 | 0.8818 | 0.5708 |
69
+ | 0.353 | 10.14 | 1500 | 0.9244 | 0.5781 |
70
+ | 0.334 | 10.81 | 1600 | 0.9009 | 0.5638 |
71
+ | 0.2917 | 11.49 | 1700 | 1.0132 | 0.5828 |
72
+ | 0.29 | 12.16 | 1800 | 0.9696 | 0.5668 |
73
+ | 0.2691 | 12.84 | 1900 | 0.9811 | 0.5455 |
74
+ | 0.25 | 13.51 | 2000 | 0.9951 | 0.5624 |
75
+ | 0.2467 | 14.19 | 2100 | 0.9653 | 0.5573 |
76
+ | 0.2242 | 14.86 | 2200 | 0.9714 | 0.5378 |
77
+ | 0.2066 | 15.54 | 2300 | 0.9829 | 0.5394 |
78
+ | 0.2075 | 16.22 | 2400 | 1.0547 | 0.5520 |
79
+ | 0.1923 | 16.89 | 2500 | 1.0014 | 0.5397 |
80
+ | 0.1919 | 17.57 | 2600 | 0.9978 | 0.5477 |
81
+ | 0.1908 | 18.24 | 2700 | 1.1064 | 0.5397 |
82
+ | 0.157 | 18.92 | 2800 | 1.0629 | 0.5238 |
83
+ | 0.159 | 19.59 | 2900 | 1.0642 | 0.5321 |
84
+ | 0.1652 | 20.27 | 3000 | 1.0207 | 0.5328 |
85
+ | 0.141 | 20.95 | 3100 | 0.9948 | 0.5312 |
86
+ | 0.1417 | 21.62 | 3200 | 1.0338 | 0.5328 |
87
+ | 0.1514 | 22.3 | 3300 | 1.0513 | 0.5313 |
88
+ | 0.1365 | 22.97 | 3400 | 1.0357 | 0.5291 |
89
+ | 0.1319 | 23.65 | 3500 | 1.0587 | 0.5167 |
90
+ | 0.1298 | 24.32 | 3600 | 1.0636 | 0.5236 |
91
+ | 0.1245 | 25.0 | 3700 | 1.1367 | 0.5280 |
92
+ | 0.1114 | 25.68 | 3800 | 1.0633 | 0.5200 |
93
+ | 0.1088 | 26.35 | 3900 | 1.0495 | 0.5210 |
94
+ | 0.1175 | 27.03 | 4000 | 1.0897 | 0.5095 |
95
+ | 0.1043 | 27.7 | 4100 | 1.0580 | 0.5309 |
96
+ | 0.0951 | 28.38 | 4200 | 1.0448 | 0.5067 |
97
+ | 0.1011 | 29.05 | 4300 | 1.0665 | 0.5137 |
98
+ | 0.0889 | 29.73 | 4400 | 1.0579 | 0.5026 |
99
+ | 0.0833 | 30.41 | 4500 | 1.0740 | 0.5037 |
100
+ | 0.0889 | 31.08 | 4600 | 1.0933 | 0.5083 |
101
+ | 0.0784 | 31.76 | 4700 | 1.0715 | 0.5089 |
102
+ | 0.0767 | 32.43 | 4800 | 1.0658 | 0.5049 |
103
+ | 0.0769 | 33.11 | 4900 | 1.1118 | 0.4979 |
104
+ | 0.0722 | 33.78 | 5000 | 1.1413 | 0.4986 |
105
+ | 0.0709 | 34.46 | 5100 | 1.0706 | 0.4885 |
106
+ | 0.0664 | 35.14 | 5200 | 1.1217 | 0.4884 |
107
+ | 0.0648 | 35.81 | 5300 | 1.1298 | 0.4941 |
108
+ | 0.0657 | 36.49 | 5400 | 1.1330 | 0.4920 |
109
+ | 0.0582 | 37.16 | 5500 | 1.0598 | 0.4835 |
110
+ | 0.0602 | 37.84 | 5600 | 1.1097 | 0.4943 |
111
+ | 0.0598 | 38.51 | 5700 | 1.0976 | 0.4876 |
112
+ | 0.0547 | 39.19 | 5800 | 1.0734 | 0.4825 |
113
+ | 0.0561 | 39.86 | 5900 | 1.0926 | 0.4850 |
114
+ | 0.0516 | 40.54 | 6000 | 1.1579 | 0.4751 |
115
+ | 0.0478 | 41.22 | 6100 | 1.1384 | 0.4706 |
116
+ | 0.0396 | 41.89 | 6200 | 1.1462 | 0.4739 |
117
+ | 0.0472 | 42.57 | 6300 | 1.1277 | 0.4732 |
118
+ | 0.0447 | 43.24 | 6400 | 1.1517 | 0.4752 |
119
+ | 0.0423 | 43.92 | 6500 | 1.1219 | 0.4784 |
120
+ | 0.0426 | 44.59 | 6600 | 1.1311 | 0.4724 |
121
+ | 0.0391 | 45.27 | 6700 | 1.1135 | 0.4692 |
122
+ | 0.0362 | 45.95 | 6800 | 1.0878 | 0.4645 |
123
+ | 0.0329 | 46.62 | 6900 | 1.1137 | 0.4668 |
124
+ | 0.0356 | 47.3 | 7000 | 1.1233 | 0.4687 |
125
+ | 0.0328 | 47.97 | 7100 | 1.1238 | 0.4653 |
126
+ | 0.0323 | 48.65 | 7200 | 1.1307 | 0.4646 |
127
+ | 0.0325 | 49.32 | 7300 | 1.1242 | 0.4645 |
128
+ | 0.03 | 50.0 | 7400 | 1.1257 | 0.4631 |
129
+
130
+
131
+ ### Framework versions
132
+
133
+ - Transformers 4.16.2
134
+ - Pytorch 1.10.0+cu111
135
+ - Datasets 1.18.3
136
+ - Tokenizers 0.11.0