Hugging Face's logo

Mihaj
/
wav2vec2-large-mms-1b-livvi-karelian-CodeSwitching

Automatic Speech Recognition
Transformers
TensorBoard
Safetensors
wav2vec2
Generated from Trainer
Inference Endpoints
Model card Files Metrics Community
Mihaj commited on
Commit
2c7c18f
·
verified ·
1 Parent(s): 89b1e79

End of training

Browse files
Files changed (2) hide show
  1. README.md +26 -106
  2. adapter.krl.safetensors +1 -1
README.md CHANGED
@@ -1,8 +1,8 @@
1
  ---
2
  license: cc-by-nc-4.0
 
3
  tags:
4
  - generated_from_trainer
5
- base_model: facebook/mms-1b-all
6
  metrics:
7
  - wer
8
  model-index:
@@ -17,9 +17,9 @@ should probably proofread and complete it, then remove this comment. -->
17
 
18
  This model is a fine-tuned version of [facebook/mms-1b-all](https://huggingface.co/facebook/mms-1b-all) on the None dataset.
19
  It achieves the following results on the evaluation set:
20
- - Loss: 0.4475
21
- - Wer: 0.4337
22
- - Cer: 0.0962
23
 
24
  ## Model description
25
 
@@ -38,7 +38,7 @@ More information needed
38
  ### Training hyperparameters
39
 
40
  The following hyperparameters were used during training:
41
- - learning_rate: 0.0007
42
  - train_batch_size: 8
43
  - eval_batch_size: 8
44
  - seed: 42
@@ -47,7 +47,7 @@ The following hyperparameters were used during training:
47
  - total_train_batch_size: 32
48
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
49
  - lr_scheduler_type: linear
50
- - lr_scheduler_warmup_steps: 100
51
  - training_steps: 10000
52
  - mixed_precision_training: Native AMP
53
 
@@ -55,106 +55,26 @@ The following hyperparameters were used during training:
55
 
56
  | Training Loss | Epoch | Step | Validation Loss | Wer | Cer |
57
  |:-------------:|:-------:|:-----:|:---------------:|:------:|:------:|
58
- | 4.5669 | 0.9070 | 100 | 0.7117 | 0.7585 | 0.1988 |
59
- | 0.687 | 1.8141 | 200 | 0.5073 | 0.6150 | 0.1476 |
60
- | 0.5978 | 2.7211 | 300 | 0.4575 | 0.5917 | 0.1373 |
61
- | 0.5536 | 3.6281 | 400 | 0.4288 | 0.5484 | 0.1297 |
62
- | 0.524 | 4.5351 | 500 | 0.4283 | 0.5639 | 0.1295 |
63
- | 0.4979 | 5.4422 | 600 | 0.4172 | 0.5474 | 0.1262 |
64
- | 0.4791 | 6.3492 | 700 | 0.4048 | 0.5311 | 0.1207 |
65
- | 0.4809 | 7.2562 | 800 | 0.4005 | 0.5217 | 0.1204 |
66
- | 0.4434 | 8.1633 | 900 | 0.3877 | 0.5122 | 0.1171 |
67
- | 0.437 | 9.0703 | 1000 | 0.3845 | 0.5105 | 0.1163 |
68
- | 0.4251 | 9.9773 | 1100 | 0.3829 | 0.5037 | 0.1146 |
69
- | 0.3866 | 10.8844 | 1200 | 0.3638 | 0.4861 | 0.1114 |
70
- | 0.3956 | 11.7914 | 1300 | 0.3726 | 0.4922 | 0.1128 |
71
- | 0.3796 | 12.6984 | 1400 | 0.3644 | 0.4922 | 0.1102 |
72
- | 0.3698 | 13.6054 | 1500 | 0.3646 | 0.4800 | 0.1091 |
73
- | 0.3657 | 14.5125 | 1600 | 0.3664 | 0.4780 | 0.1088 |
74
- | 0.3506 | 15.4195 | 1700 | 0.3609 | 0.4740 | 0.1087 |
75
- | 0.3476 | 16.3265 | 1800 | 0.3542 | 0.4770 | 0.1077 |
76
- | 0.3461 | 17.2336 | 1900 | 0.3639 | 0.4760 | 0.1070 |
77
- | 0.3194 | 18.1406 | 2000 | 0.3556 | 0.4733 | 0.1054 |
78
- | 0.3191 | 19.0476 | 2100 | 0.3639 | 0.4692 | 0.1055 |
79
- | 0.3076 | 19.9546 | 2200 | 0.3565 | 0.4614 | 0.1034 |
80
- | 0.3056 | 20.8617 | 2300 | 0.3597 | 0.4652 | 0.1047 |
81
- | 0.3068 | 21.7687 | 2400 | 0.3620 | 0.4740 | 0.1048 |
82
- | 0.2913 | 22.6757 | 2500 | 0.3535 | 0.4601 | 0.1033 |
83
- | 0.2875 | 23.5828 | 2600 | 0.3610 | 0.4672 | 0.1035 |
84
- | 0.2845 | 24.4898 | 2700 | 0.3586 | 0.4560 | 0.1020 |
85
- | 0.2793 | 25.3968 | 2800 | 0.3633 | 0.4587 | 0.1028 |
86
- | 0.273 | 26.3039 | 2900 | 0.3638 | 0.4614 | 0.1028 |
87
- | 0.2682 | 27.2109 | 3000 | 0.3725 | 0.4699 | 0.1029 |
88
- | 0.2549 | 28.1179 | 3100 | 0.3669 | 0.4540 | 0.1015 |
89
- | 0.2535 | 29.0249 | 3200 | 0.3630 | 0.4604 | 0.1026 |
90
- | 0.2464 | 29.9320 | 3300 | 0.3601 | 0.4601 | 0.1023 |
91
- | 0.2517 | 30.8390 | 3400 | 0.3680 | 0.4621 | 0.1024 |
92
- | 0.2424 | 31.7460 | 3500 | 0.3773 | 0.4597 | 0.1028 |
93
- | 0.2455 | 32.6531 | 3600 | 0.3706 | 0.4614 | 0.1019 |
94
- | 0.2311 | 33.5601 | 3700 | 0.3667 | 0.4530 | 0.0992 |
95
- | 0.2318 | 34.4671 | 3800 | 0.3750 | 0.4597 | 0.1022 |
96
- | 0.2233 | 35.3741 | 3900 | 0.3643 | 0.4662 | 0.1027 |
97
- | 0.2246 | 36.2812 | 4000 | 0.3737 | 0.4570 | 0.1013 |
98
- | 0.2187 | 37.1882 | 4100 | 0.3864 | 0.4547 | 0.1001 |
99
- | 0.2199 | 38.0952 | 4200 | 0.3925 | 0.4418 | 0.0980 |
100
- | 0.2167 | 39.0023 | 4300 | 0.3808 | 0.4533 | 0.0992 |
101
- | 0.2131 | 39.9093 | 4400 | 0.3861 | 0.4506 | 0.0996 |
102
- | 0.2044 | 40.8163 | 4500 | 0.3771 | 0.4581 | 0.1005 |
103
- | 0.1981 | 41.7234 | 4600 | 0.3759 | 0.4611 | 0.1021 |
104
- | 0.2013 | 42.6304 | 4700 | 0.3817 | 0.4641 | 0.1028 |
105
- | 0.198 | 43.5374 | 4800 | 0.3786 | 0.4479 | 0.0990 |
106
- | 0.1905 | 44.4444 | 4900 | 0.3889 | 0.4465 | 0.0976 |
107
- | 0.1829 | 45.3515 | 5000 | 0.3895 | 0.4496 | 0.0982 |
108
- | 0.1912 | 46.2585 | 5100 | 0.3894 | 0.4459 | 0.0983 |
109
- | 0.1879 | 47.1655 | 5200 | 0.3986 | 0.4486 | 0.0980 |
110
- | 0.1817 | 48.0726 | 5300 | 0.3865 | 0.4476 | 0.0996 |
111
- | 0.1803 | 48.9796 | 5400 | 0.4024 | 0.4408 | 0.0989 |
112
- | 0.1752 | 49.8866 | 5500 | 0.3871 | 0.4493 | 0.0994 |
113
- | 0.1755 | 50.7937 | 5600 | 0.3969 | 0.4442 | 0.0983 |
114
- | 0.1794 | 51.7007 | 5700 | 0.4069 | 0.4364 | 0.0969 |
115
- | 0.1739 | 52.6077 | 5800 | 0.4048 | 0.4357 | 0.0981 |
116
- | 0.1691 | 53.5147 | 5900 | 0.4082 | 0.4469 | 0.0982 |
117
- | 0.1648 | 54.4218 | 6000 | 0.4227 | 0.4422 | 0.0984 |
118
- | 0.1748 | 55.3288 | 6100 | 0.4027 | 0.4435 | 0.0990 |
119
- | 0.1668 | 56.2358 | 6200 | 0.4311 | 0.4438 | 0.0990 |
120
- | 0.1617 | 57.1429 | 6300 | 0.4116 | 0.4472 | 0.0988 |
121
- | 0.1617 | 58.0499 | 6400 | 0.4051 | 0.4432 | 0.0974 |
122
- | 0.1625 | 58.9569 | 6500 | 0.4146 | 0.4374 | 0.0975 |
123
- | 0.1594 | 59.8639 | 6600 | 0.4108 | 0.4435 | 0.0988 |
124
- | 0.1668 | 60.7710 | 6700 | 0.4095 | 0.4327 | 0.0969 |
125
- | 0.1512 | 61.6780 | 6800 | 0.4172 | 0.4378 | 0.0966 |
126
- | 0.15 | 62.5850 | 6900 | 0.4135 | 0.4422 | 0.0982 |
127
- | 0.1537 | 63.4921 | 7000 | 0.4326 | 0.4401 | 0.0980 |
128
- | 0.151 | 64.3991 | 7100 | 0.4255 | 0.4411 | 0.0980 |
129
- | 0.152 | 65.3061 | 7200 | 0.4236 | 0.4378 | 0.0976 |
130
- | 0.1468 | 66.2132 | 7300 | 0.4207 | 0.4526 | 0.0996 |
131
- | 0.1497 | 67.1202 | 7400 | 0.4160 | 0.4398 | 0.0980 |
132
- | 0.1509 | 68.0272 | 7500 | 0.4210 | 0.4401 | 0.0975 |
133
- | 0.138 | 68.9342 | 7600 | 0.4224 | 0.4469 | 0.0984 |
134
- | 0.1435 | 69.8413 | 7700 | 0.4217 | 0.4425 | 0.0968 |
135
- | 0.1465 | 70.7483 | 7800 | 0.4280 | 0.4354 | 0.0964 |
136
- | 0.1367 | 71.6553 | 7900 | 0.4260 | 0.4405 | 0.0978 |
137
- | 0.1445 | 72.5624 | 8000 | 0.4253 | 0.4394 | 0.0980 |
138
- | 0.1379 | 73.4694 | 8100 | 0.4250 | 0.4347 | 0.0968 |
139
- | 0.1385 | 74.3764 | 8200 | 0.4303 | 0.4391 | 0.0975 |
140
- | 0.1334 | 75.2834 | 8300 | 0.4377 | 0.4357 | 0.0968 |
141
- | 0.1372 | 76.1905 | 8400 | 0.4423 | 0.4323 | 0.0957 |
142
- | 0.1329 | 77.0975 | 8500 | 0.4281 | 0.4334 | 0.0967 |
143
- | 0.1322 | 78.0045 | 8600 | 0.4366 | 0.4313 | 0.0956 |
144
- | 0.1327 | 78.9116 | 8700 | 0.4377 | 0.4344 | 0.0960 |
145
- | 0.1313 | 79.8186 | 8800 | 0.4405 | 0.4361 | 0.0964 |
146
- | 0.1315 | 80.7256 | 8900 | 0.4401 | 0.4350 | 0.0959 |
147
- | 0.134 | 81.6327 | 9000 | 0.4370 | 0.4364 | 0.0963 |
148
- | 0.1351 | 82.5397 | 9100 | 0.4389 | 0.4364 | 0.0962 |
149
- | 0.1289 | 83.4467 | 9200 | 0.4504 | 0.4384 | 0.0972 |
150
- | 0.1308 | 84.3537 | 9300 | 0.4447 | 0.4334 | 0.0963 |
151
- | 0.1321 | 85.2608 | 9400 | 0.4509 | 0.4334 | 0.0965 |
152
- | 0.1296 | 86.1678 | 9500 | 0.4489 | 0.4323 | 0.0960 |
153
- | 0.1277 | 87.0748 | 9600 | 0.4489 | 0.4374 | 0.0963 |
154
- | 0.1245 | 87.9819 | 9700 | 0.4475 | 0.4344 | 0.0965 |
155
- | 0.1267 | 88.8889 | 9800 | 0.4459 | 0.4323 | 0.0952 |
156
- | 0.1259 | 89.7959 | 9900 | 0.4457 | 0.4320 | 0.0959 |
157
- | 0.1247 | 90.7029 | 10000 | 0.4475 | 0.4337 | 0.0962 |
158
 
159
 
160
  ### Framework versions
 
1
  ---
2
  license: cc-by-nc-4.0
3
+ base_model: facebook/mms-1b-all
4
  tags:
5
  - generated_from_trainer
 
6
  metrics:
7
  - wer
8
  model-index:
 
17
 
18
  This model is a fine-tuned version of [facebook/mms-1b-all](https://huggingface.co/facebook/mms-1b-all) on the None dataset.
19
  It achieves the following results on the evaluation set:
20
+ - Loss: 0.3113
21
+ - Wer: 0.4087
22
+ - Cer: 0.0910
23
 
24
  ## Model description
25
 
 
38
  ### Training hyperparameters
39
 
40
  The following hyperparameters were used during training:
41
+ - learning_rate: 0.0005
42
  - train_batch_size: 8
43
  - eval_batch_size: 8
44
  - seed: 42
 
47
  - total_train_batch_size: 32
48
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
49
  - lr_scheduler_type: linear
50
+ - lr_scheduler_warmup_steps: 500
51
  - training_steps: 10000
52
  - mixed_precision_training: Native AMP
53
 
 
55
 
56
  | Training Loss | Epoch | Step | Validation Loss | Wer | Cer |
57
  |:-------------:|:-------:|:-----:|:---------------:|:------:|:------:|
58
+ | 1.4219 | 4.5351 | 500 | 0.4570 | 0.5677 | 0.1335 |
59
+ | 0.5951 | 9.0703 | 1000 | 0.4008 | 0.5142 | 0.1186 |
60
+ | 0.5314 | 13.6054 | 1500 | 0.3725 | 0.4942 | 0.1126 |
61
+ | 0.4916 | 18.1406 | 2000 | 0.3626 | 0.4692 | 0.1067 |
62
+ | 0.4563 | 22.6757 | 2500 | 0.3465 | 0.4540 | 0.1035 |
63
+ | 0.4331 | 27.2109 | 3000 | 0.3310 | 0.4455 | 0.1010 |
64
+ | 0.4129 | 31.7460 | 3500 | 0.3283 | 0.4516 | 0.1019 |
65
+ | 0.394 | 36.2812 | 4000 | 0.3289 | 0.4482 | 0.0994 |
66
+ | 0.3715 | 40.8163 | 4500 | 0.3203 | 0.4374 | 0.0985 |
67
+ | 0.3646 | 45.3515 | 5000 | 0.3109 | 0.4327 | 0.0966 |
68
+ | 0.3508 | 49.8866 | 5500 | 0.3136 | 0.4276 | 0.0958 |
69
+ | 0.3376 | 54.4218 | 6000 | 0.3198 | 0.4246 | 0.0950 |
70
+ | 0.3283 | 58.9569 | 6500 | 0.3203 | 0.4232 | 0.0943 |
71
+ | 0.3222 | 63.4921 | 7000 | 0.3126 | 0.4134 | 0.0932 |
72
+ | 0.3104 | 68.0272 | 7500 | 0.3140 | 0.4168 | 0.0933 |
73
+ | 0.3026 | 72.5624 | 8000 | 0.3136 | 0.4110 | 0.0920 |
74
+ | 0.3003 | 77.0975 | 8500 | 0.3137 | 0.4175 | 0.0926 |
75
+ | 0.2896 | 81.6327 | 9000 | 0.3150 | 0.4107 | 0.0912 |
76
+ | 0.2885 | 86.1678 | 9500 | 0.3110 | 0.4090 | 0.0914 |
77
+ | 0.2869 | 90.7029 | 10000 | 0.3113 | 0.4087 | 0.0910 |
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
78
 
79
 
80
  ### Framework versions
adapter.krl.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:68d475510de03db55e74a2884ad85b68537106925cb3e66b607201b7d89c92cd
3
  size 4424388
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:bacaee74e07b9cbb26b5c7385b643f0e20f9ecafe2e059435ba30ff1dce4c3e4
3
  size 4424388