Hugging Face's logo

ctaguchi
/
ssc-ukv-mms-model-mix-adapt-max3-devtrain

Automatic Speech Recognition
Transformers
Safetensors
wav2vec2
Generated from Trainer
Model card Files
xet
 Community
ctaguchi commited on
Commit
d5654f7
·
verified ·
1 Parent(s): e8a04a5

End of training

Browse files
Files changed (4) hide show
  1. README.md +27 -71
  2. adapter.ukv.safetensors +1 -1
  3. model.safetensors +1 -1
  4. training_args.bin +1 -1
README.md CHANGED
@@ -1,7 +1,5 @@
1
  ---
2
  library_name: transformers
3
- license: cc-by-nc-4.0
4
- base_model: facebook/mms-1b-all
5
  tags:
6
  - generated_from_trainer
7
  metrics:
@@ -16,11 +14,11 @@ should probably proofread and complete it, then remove this comment. -->
16
 
17
  # ssc-ukv-mms-model-mix-adapt-max3-devtrain
18
 
19
- This model is a fine-tuned version of [facebook/mms-1b-all](https://huggingface.co/facebook/mms-1b-all) on an unknown dataset.
20
  It achieves the following results on the evaluation set:
21
- - Loss: 0.4174
22
- - Cer: 0.1226
23
- - Wer: 0.3697
24
 
25
  ## Model description
26
 
@@ -40,81 +38,39 @@ More information needed
40
 
41
  The following hyperparameters were used during training:
42
  - learning_rate: 0.0005
43
- - train_batch_size: 8
44
  - eval_batch_size: 6
45
  - seed: 42
46
  - gradient_accumulation_steps: 2
47
- - total_train_batch_size: 16
48
  - optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
49
  - lr_scheduler_type: linear
50
  - lr_scheduler_warmup_steps: 100
51
- - num_epochs: 20
52
  - mixed_precision_training: Native AMP
53
 
54
  ### Training results
55
 
56
- | Training Loss | Epoch | Step | Validation Loss | Cer | Wer |
57
- |:-------------:|:-------:|:-----:|:---------------:|:------:|:------:|
58
- | 1.2437 | 0.3289 | 200 | 0.7379 | 0.1992 | 0.6945 |
59
- | 1.0434 | 0.6579 | 400 | 0.6090 | 0.1589 | 0.5133 |
60
- | 0.9835 | 0.9868 | 600 | 0.5756 | 0.1474 | 0.4771 |
61
- | 0.9662 | 1.3158 | 800 | 0.5705 | 0.1542 | 0.5116 |
62
- | 0.9153 | 1.6447 | 1000 | 0.5622 | 0.1487 | 0.4749 |
63
- | 0.9532 | 1.9737 | 1200 | 0.5581 | 0.1427 | 0.4555 |
64
- | 0.892 | 2.3026 | 1400 | 0.5391 | 0.1414 | 0.4459 |
65
- | 0.9139 | 2.6316 | 1600 | 0.5316 | 0.1391 | 0.4401 |
66
- | 0.8517 | 2.9605 | 1800 | 0.5258 | 0.1360 | 0.4278 |
67
- | 0.8843 | 3.2895 | 2000 | 0.5211 | 0.1368 | 0.4300 |
68
- | 0.8701 | 3.6184 | 2200 | 0.5098 | 0.1344 | 0.4200 |
69
- | 0.8829 | 3.9474 | 2400 | 0.5076 | 0.1352 | 0.4289 |
70
- | 0.8826 | 4.2763 | 2600 | 0.5064 | 0.1304 | 0.4107 |
71
- | 0.8749 | 4.6053 | 2800 | 0.4948 | 0.1330 | 0.4164 |
72
- | 0.8575 | 4.9342 | 3000 | 0.4896 | 0.1323 | 0.4170 |
73
- | 0.8163 | 5.2632 | 3200 | 0.4875 | 0.1313 | 0.4102 |
74
- | 0.8357 | 5.5921 | 3400 | 0.4869 | 0.1309 | 0.4094 |
75
- | 0.8344 | 5.9211 | 3600 | 0.4786 | 0.1289 | 0.4026 |
76
- | 0.7891 | 6.25 | 3800 | 0.4833 | 0.1284 | 0.3991 |
77
- | 0.7968 | 6.5789 | 4000 | 0.4777 | 0.1298 | 0.4092 |
78
- | 0.8714 | 6.9079 | 4200 | 0.4725 | 0.1292 | 0.4085 |
79
- | 0.8048 | 7.2368 | 4400 | 0.4730 | 0.1272 | 0.3987 |
80
- | 0.8177 | 7.5658 | 4600 | 0.4768 | 0.1276 | 0.4035 |
81
- | 0.7767 | 7.8947 | 4800 | 0.4747 | 0.1264 | 0.3899 |
82
- | 0.8006 | 8.2237 | 5000 | 0.4752 | 0.1269 | 0.3967 |
83
- | 0.778 | 8.5526 | 5200 | 0.4616 | 0.1248 | 0.3886 |
84
- | 0.7914 | 8.8816 | 5400 | 0.4602 | 0.1261 | 0.3848 |
85
- | 0.7422 | 9.2105 | 5600 | 0.4631 | 0.1255 | 0.3861 |
86
- | 0.7651 | 9.5395 | 5800 | 0.4665 | 0.1275 | 0.3910 |
87
- | 0.7718 | 9.8684 | 6000 | 0.4504 | 0.1258 | 0.3866 |
88
- | 0.7611 | 10.1974 | 6200 | 0.4460 | 0.1265 | 0.3907 |
89
- | 0.7488 | 10.5263 | 6400 | 0.4434 | 0.1274 | 0.3893 |
90
- | 0.7591 | 10.8553 | 6600 | 0.4497 | 0.1276 | 0.3867 |
91
- | 0.747 | 11.1842 | 6800 | 0.4447 | 0.1262 | 0.3855 |
92
- | 0.7428 | 11.5132 | 7000 | 0.4442 | 0.1263 | 0.3825 |
93
- | 0.7817 | 11.8421 | 7200 | 0.4391 | 0.1269 | 0.3862 |
94
- | 0.7186 | 12.1711 | 7400 | 0.4461 | 0.1288 | 0.3916 |
95
- | 0.7502 | 12.5 | 7600 | 0.4350 | 0.1272 | 0.3830 |
96
- | 0.7368 | 12.8289 | 7800 | 0.4425 | 0.1272 | 0.3863 |
97
- | 0.73 | 13.1579 | 8000 | 0.4354 | 0.1271 | 0.3858 |
98
- | 0.7381 | 13.4868 | 8200 | 0.4461 | 0.1258 | 0.3824 |
99
- | 0.7319 | 13.8158 | 8400 | 0.4418 | 0.1260 | 0.3791 |
100
- | 0.6887 | 14.1447 | 8600 | 0.4373 | 0.1264 | 0.3824 |
101
- | 0.7257 | 14.4737 | 8800 | 0.4336 | 0.1250 | 0.3765 |
102
- | 0.7562 | 14.8026 | 9000 | 0.4317 | 0.1249 | 0.3765 |
103
- | 0.7387 | 15.1316 | 9200 | 0.4401 | 0.1260 | 0.3829 |
104
- | 0.6813 | 15.4605 | 9400 | 0.4305 | 0.1251 | 0.3825 |
105
- | 0.7144 | 15.7895 | 9600 | 0.4270 | 0.1238 | 0.3760 |
106
- | 0.7222 | 16.1184 | 9800 | 0.4268 | 0.1246 | 0.3783 |
107
- | 0.7173 | 16.4474 | 10000 | 0.4295 | 0.1239 | 0.3737 |
108
- | 0.6838 | 16.7763 | 10200 | 0.4264 | 0.1243 | 0.3768 |
109
- | 0.6975 | 17.1053 | 10400 | 0.4241 | 0.1241 | 0.3760 |
110
- | 0.6996 | 17.4342 | 10600 | 0.4221 | 0.1228 | 0.3706 |
111
- | 0.7098 | 17.7632 | 10800 | 0.4228 | 0.1235 | 0.3741 |
112
- | 0.7402 | 18.0921 | 11000 | 0.4222 | 0.1220 | 0.3708 |
113
- | 0.6887 | 18.4211 | 11200 | 0.4213 | 0.1238 | 0.3747 |
114
- | 0.7038 | 18.75 | 11400 | 0.4183 | 0.1227 | 0.3707 |
115
- | 0.6955 | 19.0789 | 11600 | 0.4221 | 0.1224 | 0.3706 |
116
- | 0.7031 | 19.4079 | 11800 | 0.4200 | 0.1228 | 0.3715 |
117
- | 0.6791 | 19.7368 | 12000 | 0.4174 | 0.1226 | 0.3697 |
118
 
119
 
120
  ### Framework versions
 
1
  ---
2
  library_name: transformers
 
 
3
  tags:
4
  - generated_from_trainer
5
  metrics:
 
14
 
15
  # ssc-ukv-mms-model-mix-adapt-max3-devtrain
16
 
17
+ This model was trained from scratch on an unknown dataset.
18
  It achieves the following results on the evaluation set:
19
+ - Loss: 0.4336
20
+ - Cer: 0.1233
21
+ - Wer: 0.3768
22
 
23
  ## Model description
24
 
 
38
 
39
  The following hyperparameters were used during training:
40
  - learning_rate: 0.0005
41
+ - train_batch_size: 1
42
  - eval_batch_size: 6
43
  - seed: 42
44
  - gradient_accumulation_steps: 2
45
+ - total_train_batch_size: 2
46
  - optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
47
  - lr_scheduler_type: linear
48
  - lr_scheduler_warmup_steps: 100
49
+ - num_epochs: 5
50
  - mixed_precision_training: Native AMP
51
 
52
  ### Training results
53
 
54
+ | Training Loss | Epoch | Step | Validation Loss | Cer | Wer |
55
+ |:-------------:|:------:|:----:|:---------------:|:------:|:------:|
56
+ | 0.5931 | 0.2658 | 200 | 0.4546 | 0.1241 | 0.3862 |
57
+ | 0.5973 | 0.5316 | 400 | 0.4389 | 0.1270 | 0.3861 |
58
+ | 0.523 | 0.7973 | 600 | 0.4552 | 0.1265 | 0.3881 |
59
+ | 0.5222 | 1.0625 | 800 | 0.4621 | 0.1277 | 0.3882 |
60
+ | 0.5543 | 1.3282 | 1000 | 0.4657 | 0.1292 | 0.3916 |
61
+ | 0.519 | 1.5940 | 1200 | 0.4536 | 0.1237 | 0.3868 |
62
+ | 0.5157 | 1.8598 | 1400 | 0.4498 | 0.1230 | 0.3805 |
63
+ | 0.4708 | 2.1249 | 1600 | 0.4514 | 0.1308 | 0.4045 |
64
+ | 0.4592 | 2.3907 | 1800 | 0.4524 | 0.1235 | 0.3783 |
65
+ | 0.5358 | 2.6565 | 2000 | 0.4444 | 0.1233 | 0.3768 |
66
+ | 0.4886 | 2.9223 | 2200 | 0.4438 | 0.1237 | 0.3836 |
67
+ | 0.4571 | 3.1874 | 2400 | 0.4377 | 0.1239 | 0.3858 |
68
+ | 0.441 | 3.4532 | 2600 | 0.4380 | 0.1244 | 0.3836 |
69
+ | 0.4925 | 3.7189 | 2800 | 0.4426 | 0.1236 | 0.3834 |
70
+ | 0.4775 | 3.9847 | 3000 | 0.4340 | 0.1235 | 0.3814 |
71
+ | 0.4715 | 4.2498 | 3200 | 0.4367 | 0.1227 | 0.3800 |
72
+ | 0.4978 | 4.5156 | 3400 | 0.4337 | 0.1237 | 0.3803 |
73
+ | 0.4331 | 4.7814 | 3600 | 0.4336 | 0.1233 | 0.3768 |
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
74
 
75
 
76
  ### Framework versions
adapter.ukv.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:5f65c29e60b3f53b287cd961ef3ac1db3aa2201fb9dafd5c0f4d790ea46ad998
3
  size 8993260
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:86a6a8f75efd34b98021ed7a7c9947d7786d8f47792a40559f08d832dadd1ddf
3
  size 8993260
model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:c0f27c73fe03e38e48456549c590b0861be28e464796390020b381d63bc4b95b
3
  size 3859085636
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8ff316bd65f0ad9aeffff14cebb0a3176d2001e6f897a0a369df29619d467fb9
3
  size 3859085636
training_args.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:9d3d0ba5c25ac02a16823f5c4f09f598977392c7e1c41079b6862081c1eff585
3
  size 5969
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7bfbfc1851356c49b6ad4df043e9d9961d401d1e672e59beb5bb008e60881233
3
  size 5969