Hugging Face's logo

nrshoudi
/
wav2vec2-large-xls-r-300m-Arabic

Automatic Speech Recognition
Transformers
PyTorch
wav2vec2
Generated from Trainer
Inference Endpoints
Model card Files Community
nrshoudi commited on
Commit
d9d2771
•
1 Parent(s): 7838392

update model card README.md

Browse files
Files changed (1) hide show
  1. README.md +33 -33
README.md CHANGED
@@ -16,9 +16,9 @@ should probably proofread and complete it, then remove this comment. -->
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 None dataset.
18
  It achieves the following results on the evaluation set:
19
- - Loss: 0.0085
20
- - Wer: 0.0152
21
- - Cer: 0.0043
22
 
23
  ## Model description
24
 
@@ -53,36 +53,36 @@ The following hyperparameters were used during training:
53
 
54
  | Training Loss | Epoch | Step | Validation Loss | Wer | Cer |
55
  |:-------------:|:-----:|:----:|:---------------:|:------:|:------:|
56
- | 17.5605 | 1.0 | 51 | 4.7252 | 1.0 | 1.0 |
57
- | 3.9344 | 2.0 | 102 | 3.1278 | 1.0 | 1.0 |
58
- | 3.1764 | 3.0 | 153 | 3.1129 | 1.0 | 1.0 |
59
- | 3.1404 | 4.0 | 204 | 3.0663 | 1.0 | 1.0 |
60
- | 3.1058 | 5.0 | 255 | 3.0273 | 1.0 | 1.0 |
61
- | 3.022 | 6.0 | 306 | 2.9014 | 1.0 | 1.0 |
62
- | 2.8023 | 7.0 | 357 | 2.4832 | 1.0 | 1.0 |
63
- | 2.2454 | 8.0 | 408 | 1.6200 | 0.9948 | 0.8759 |
64
- | 1.2981 | 9.0 | 459 | 0.4455 | 0.6949 | 0.2470 |
65
- | 0.4965 | 10.0 | 510 | 0.1542 | 0.3887 | 0.0858 |
66
- | 0.2256 | 11.0 | 561 | 0.0531 | 0.0721 | 0.0186 |
67
- | 0.1319 | 12.0 | 612 | 0.0264 | 0.0505 | 0.0137 |
68
- | 0.0905 | 13.0 | 663 | 0.0402 | 0.0271 | 0.0075 |
69
- | 0.0918 | 14.0 | 714 | 0.0348 | 0.0569 | 0.0189 |
70
- | 0.0674 | 15.0 | 765 | 0.0211 | 0.0301 | 0.0063 |
71
- | 0.0514 | 16.0 | 816 | 0.0142 | 0.0230 | 0.0065 |
72
- | 0.0428 | 17.0 | 867 | 0.0126 | 0.0182 | 0.0042 |
73
- | 0.035 | 18.0 | 918 | 0.0156 | 0.0290 | 0.0083 |
74
- | 0.0335 | 19.0 | 969 | 0.0105 | 0.0190 | 0.0059 |
75
- | 0.0307 | 20.0 | 1020 | 0.0100 | 0.0145 | 0.0041 |
76
- | 0.0224 | 21.0 | 1071 | 0.0116 | 0.0234 | 0.0083 |
77
- | 0.0231 | 22.0 | 1122 | 0.0124 | 0.0256 | 0.0079 |
78
- | 0.0222 | 23.0 | 1173 | 0.0134 | 0.0115 | 0.0036 |
79
- | 0.0195 | 24.0 | 1224 | 0.0104 | 0.0123 | 0.0029 |
80
- | 0.0201 | 25.0 | 1275 | 0.0101 | 0.0111 | 0.0025 |
81
- | 0.0187 | 26.0 | 1326 | 0.0082 | 0.0123 | 0.0028 |
82
- | 0.0169 | 27.0 | 1377 | 0.0079 | 0.0190 | 0.0056 |
83
- | 0.0141 | 28.0 | 1428 | 0.0099 | 0.0171 | 0.0047 |
84
- | 0.0157 | 29.0 | 1479 | 0.0089 | 0.0171 | 0.0047 |
85
- | 0.0123 | 30.0 | 1530 | 0.0085 | 0.0152 | 0.0043 |
86
 
87
 
88
  ### Framework versions
 
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 None dataset.
18
  It achieves the following results on the evaluation set:
19
+ - Loss: 0.0139
20
+ - Wer: 0.0196
21
+ - Cer: 0.0059
22
 
23
  ## Model description
24
 
 
53
 
54
  | Training Loss | Epoch | Step | Validation Loss | Wer | Cer |
55
  |:-------------:|:-----:|:----:|:---------------:|:------:|:------:|
56
+ | 0.0206 | 1.0 | 51 | 0.0297 | 0.0262 | 0.0063 |
57
+ | 0.0216 | 2.0 | 102 | 0.0300 | 0.0458 | 0.0106 |
58
+ | 0.0367 | 3.0 | 153 | 0.0520 | 0.0768 | 0.0189 |
59
+ | 0.0631 | 4.0 | 204 | 0.0443 | 0.0605 | 0.0140 |
60
+ | 0.0581 | 5.0 | 255 | 0.0274 | 0.0387 | 0.0099 |
61
+ | 0.0464 | 6.0 | 306 | 0.0248 | 0.0358 | 0.0091 |
62
+ | 0.032 | 7.0 | 357 | 0.0279 | 0.0317 | 0.0073 |
63
+ | 0.0351 | 8.0 | 408 | 0.0237 | 0.0299 | 0.0075 |
64
+ | 0.0286 | 9.0 | 459 | 0.0188 | 0.0494 | 0.0160 |
65
+ | 0.0191 | 10.0 | 510 | 0.0184 | 0.0428 | 0.0105 |
66
+ | 0.0217 | 11.0 | 561 | 0.0245 | 0.0244 | 0.0065 |
67
+ | 0.019 | 12.0 | 612 | 0.0190 | 0.0218 | 0.0052 |
68
+ | 0.0127 | 13.0 | 663 | 0.0183 | 0.0240 | 0.0057 |
69
+ | 0.0212 | 14.0 | 714 | 0.0174 | 0.0351 | 0.0114 |
70
+ | 0.0137 | 15.0 | 765 | 0.0196 | 0.0325 | 0.0083 |
71
+ | 0.0133 | 16.0 | 816 | 0.0204 | 0.0203 | 0.0052 |
72
+ | 0.011 | 17.0 | 867 | 0.0155 | 0.0218 | 0.0063 |
73
+ | 0.0096 | 18.0 | 918 | 0.0123 | 0.0229 | 0.0076 |
74
+ | 0.0093 | 19.0 | 969 | 0.0127 | 0.0273 | 0.0088 |
75
+ | 0.0094 | 20.0 | 1020 | 0.0138 | 0.0210 | 0.0059 |
76
+ | 0.0079 | 21.0 | 1071 | 0.0186 | 0.0225 | 0.0062 |
77
+ | 0.0082 | 22.0 | 1122 | 0.0141 | 0.0225 | 0.0061 |
78
+ | 0.0075 | 23.0 | 1173 | 0.0132 | 0.0185 | 0.0050 |
79
+ | 0.0073 | 24.0 | 1224 | 0.0137 | 0.0196 | 0.0058 |
80
+ | 0.0052 | 25.0 | 1275 | 0.0149 | 0.0188 | 0.0055 |
81
+ | 0.0059 | 26.0 | 1326 | 0.0152 | 0.0166 | 0.0055 |
82
+ | 0.0059 | 27.0 | 1377 | 0.0146 | 0.0185 | 0.0060 |
83
+ | 0.0049 | 28.0 | 1428 | 0.0138 | 0.0192 | 0.0059 |
84
+ | 0.0049 | 29.0 | 1479 | 0.0139 | 0.0196 | 0.0059 |
85
+ | 0.0045 | 30.0 | 1530 | 0.0139 | 0.0196 | 0.0059 |
86
 
87
 
88
  ### Framework versions