Hugging Face's logo

8688chris
/
Helldivers2ASR_V3

Safetensors
wav2vec2
Generated from Trainer
Model card Files
xet
Community
8688chris commited on
Commit
2e5f961
·
verified ·
1 Parent(s): 3fe066a

Model save

Browse files
Files changed (4) hide show
  1. README.md +67 -37
  2. config.json +4 -4
  3. model.safetensors +1 -1
  4. training_args.bin +2 -2
README.md CHANGED
@@ -17,8 +17,8 @@ should probably proofread and complete it, then remove this comment. -->
17
 
18
  This model is a fine-tuned version of [facebook/wav2vec2-base-960h](https://huggingface.co/facebook/wav2vec2-base-960h) on an unknown dataset.
19
  It achieves the following results on the evaluation set:
20
- - Loss: 135.5781
21
- - Wer: 0.0427
22
 
23
  ## Model description
24
 
@@ -37,49 +37,79 @@ More information needed
37
  ### Training hyperparameters
38
 
39
  The following hyperparameters were used during training:
40
- - learning_rate: 8e-05
41
  - train_batch_size: 32
42
  - eval_batch_size: 32
43
  - seed: 42
44
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
45
- - lr_scheduler_type: constant
46
- - num_epochs: 30
47
  - mixed_precision_training: Native AMP
48
 
49
  ### Training results
50
 
51
- | Training Loss | Epoch | Step | Validation Loss | Wer |
52
- |:-------------:|:-----:|:----:|:---------------:|:------:|
53
- | 1518.9794 | 1.0 | 168 | 1000.8097 | 0.4552 |
54
- | 1028.341 | 2.0 | 336 | 779.6321 | 0.3719 |
55
- | 820.663 | 3.0 | 504 | 659.3882 | 0.2988 |
56
- | 718.5459 | 4.0 | 672 | 516.9663 | 0.2303 |
57
- | 606.0076 | 5.0 | 840 | 421.7630 | 0.1998 |
58
- | 535.3864 | 6.0 | 1008 | 478.5203 | 0.2051 |
59
- | 466.5894 | 7.0 | 1176 | 440.6602 | 0.1739 |
60
- | 432.4227 | 8.0 | 1344 | 294.9408 | 0.1323 |
61
- | 373.2972 | 9.0 | 1512 | 261.7947 | 0.1122 |
62
- | 355.762 | 10.0 | 1680 | 315.7706 | 0.1300 |
63
- | 314.1884 | 11.0 | 1848 | 350.8355 | 0.1181 |
64
- | 288.5761 | 12.0 | 2016 | 310.4185 | 0.1201 |
65
- | 291.121 | 13.0 | 2184 | 255.1681 | 0.0876 |
66
- | 236.6435 | 14.0 | 2352 | 240.4549 | 0.0873 |
67
- | 219.9664 | 15.0 | 2520 | 237.7248 | 0.0922 |
68
- | 201.4798 | 16.0 | 2688 | 162.6640 | 0.0619 |
69
- | 199.2839 | 17.0 | 2856 | 232.3585 | 0.0800 |
70
- | 194.1537 | 18.0 | 3024 | 215.3707 | 0.0772 |
71
- | 177.814 | 19.0 | 3192 | 171.7732 | 0.0589 |
72
- | 166.7409 | 20.0 | 3360 | 166.2487 | 0.0597 |
73
- | 169.7996 | 21.0 | 3528 | 135.5038 | 0.0546 |
74
- | 153.8049 | 22.0 | 3696 | 150.6883 | 0.0518 |
75
- | 143.4673 | 23.0 | 3864 | 179.0132 | 0.0541 |
76
- | 147.514 | 24.0 | 4032 | 131.5579 | 0.0419 |
77
- | 138.0108 | 25.0 | 4200 | 154.8247 | 0.0493 |
78
- | 142.9634 | 26.0 | 4368 | 165.7421 | 0.0586 |
79
- | 123.1378 | 27.0 | 4536 | 160.7584 | 0.0472 |
80
- | 129.9836 | 28.0 | 4704 | 104.9703 | 0.0366 |
81
- | 113.1207 | 29.0 | 4872 | 172.9598 | 0.0490 |
82
- | 110.3937 | 30.0 | 5040 | 135.5781 | 0.0427 |
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
83
 
84
 
85
  ### Framework versions
 
17
 
18
  This model is a fine-tuned version of [facebook/wav2vec2-base-960h](https://huggingface.co/facebook/wav2vec2-base-960h) on an unknown dataset.
19
  It achieves the following results on the evaluation set:
20
+ - Loss: 217.4995
21
+ - Wer: 0.0488
22
 
23
  ## Model description
24
 
 
37
  ### Training hyperparameters
38
 
39
  The following hyperparameters were used during training:
40
+ - learning_rate: 0.0001
41
  - train_batch_size: 32
42
  - eval_batch_size: 32
43
  - seed: 42
44
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
45
+ - lr_scheduler_type: cosine
46
+ - num_epochs: 60
47
  - mixed_precision_training: Native AMP
48
 
49
  ### Training results
50
 
51
+ | Training Loss | Epoch | Step | Validation Loss | Wer |
52
+ |:-------------:|:-----:|:-----:|:---------------:|:------:|
53
+ | 1472.3185 | 1.0 | 220 | 996.5219 | 0.4717 |
54
+ | 988.7808 | 2.0 | 440 | 742.4960 | 0.3192 |
55
+ | 758.6794 | 3.0 | 660 | 726.9413 | 0.2945 |
56
+ | 601.6771 | 4.0 | 880 | 525.1385 | 0.2309 |
57
+ | 517.6635 | 5.0 | 1100 | 579.2085 | 0.2188 |
58
+ | 454.2735 | 6.0 | 1320 | 436.7187 | 0.1770 |
59
+ | 374.8431 | 7.0 | 1540 | 509.1526 | 0.1835 |
60
+ | 320.3781 | 8.0 | 1760 | 405.3188 | 0.1558 |
61
+ | 305.5322 | 9.0 | 1980 | 305.2030 | 0.1185 |
62
+ | 295.8101 | 10.0 | 2200 | 348.2143 | 0.1369 |
63
+ | 252.8921 | 11.0 | 2420 | 295.6262 | 0.1027 |
64
+ | 225.9937 | 12.0 | 2640 | 338.8898 | 0.1057 |
65
+ | 186.0668 | 13.0 | 2860 | 188.0684 | 0.0760 |
66
+ | 216.4071 | 14.0 | 3080 | 297.2408 | 0.0911 |
67
+ | 188.2282 | 15.0 | 3300 | 297.4695 | 0.1007 |
68
+ | 180.0134 | 16.0 | 3520 | 319.2300 | 0.0958 |
69
+ | 158.7572 | 17.0 | 3740 | 231.6492 | 0.0735 |
70
+ | 151.4667 | 18.0 | 3960 | 390.5204 | 0.1033 |
71
+ | 134.8128 | 19.0 | 4180 | 302.3186 | 0.0873 |
72
+ | 128.7475 | 20.0 | 4400 | 254.5555 | 0.0792 |
73
+ | 143.8669 | 21.0 | 4620 | 292.1411 | 0.0843 |
74
+ | 103.34 | 22.0 | 4840 | 418.9893 | 0.0903 |
75
+ | 122.2672 | 23.0 | 5060 | 364.5801 | 0.0954 |
76
+ | 97.266 | 24.0 | 5280 | 285.9376 | 0.0756 |
77
+ | 95.7884 | 25.0 | 5500 | 171.5240 | 0.0543 |
78
+ | 103.1276 | 26.0 | 5720 | 261.3536 | 0.0662 |
79
+ | 78.8389 | 27.0 | 5940 | 293.9692 | 0.0776 |
80
+ | 92.2303 | 28.0 | 6160 | 332.2821 | 0.0717 |
81
+ | 81.0719 | 29.0 | 6380 | 283.8300 | 0.0640 |
82
+ | 81.7117 | 30.0 | 6600 | 194.0315 | 0.0512 |
83
+ | 82.8297 | 31.0 | 6820 | 268.2238 | 0.0624 |
84
+ | 66.1784 | 32.0 | 7040 | 172.1759 | 0.0429 |
85
+ | 85.3869 | 33.0 | 7260 | 218.6479 | 0.0585 |
86
+ | 66.2284 | 34.0 | 7480 | 249.1338 | 0.0596 |
87
+ | 65.4866 | 35.0 | 7700 | 193.9302 | 0.0478 |
88
+ | 59.2717 | 36.0 | 7920 | 178.9640 | 0.0464 |
89
+ | 57.6191 | 37.0 | 8140 | 177.6202 | 0.0480 |
90
+ | 50.2033 | 38.0 | 8360 | 174.2884 | 0.0405 |
91
+ | 62.3403 | 39.0 | 8580 | 222.0836 | 0.0519 |
92
+ | 64.9 | 40.0 | 8800 | 184.1204 | 0.0433 |
93
+ | 48.9848 | 41.0 | 9020 | 159.7262 | 0.0399 |
94
+ | 53.4291 | 42.0 | 9240 | 169.9478 | 0.0393 |
95
+ | 64.6475 | 43.0 | 9460 | 214.1101 | 0.0531 |
96
+ | 56.4478 | 44.0 | 9680 | 176.6035 | 0.0458 |
97
+ | 46.7233 | 45.0 | 9900 | 224.5517 | 0.0504 |
98
+ | 49.9667 | 46.0 | 10120 | 190.4445 | 0.0472 |
99
+ | 46.9387 | 47.0 | 10340 | 206.6020 | 0.0486 |
100
+ | 42.602 | 48.0 | 10560 | 196.2499 | 0.0470 |
101
+ | 42.0918 | 49.0 | 10780 | 177.2713 | 0.0460 |
102
+ | 44.1689 | 50.0 | 11000 | 186.2561 | 0.0466 |
103
+ | 41.0818 | 51.0 | 11220 | 212.5681 | 0.0508 |
104
+ | 37.013 | 52.0 | 11440 | 207.1802 | 0.0484 |
105
+ | 36.6795 | 53.0 | 11660 | 205.6143 | 0.0468 |
106
+ | 49.4436 | 54.0 | 11880 | 210.2148 | 0.0486 |
107
+ | 47.5393 | 55.0 | 12100 | 216.7866 | 0.0498 |
108
+ | 47.0014 | 56.0 | 12320 | 216.6146 | 0.0490 |
109
+ | 52.2467 | 57.0 | 12540 | 218.1209 | 0.0494 |
110
+ | 51.5985 | 58.0 | 12760 | 217.3818 | 0.0490 |
111
+ | 51.9654 | 59.0 | 12980 | 217.6073 | 0.0490 |
112
+ | 40.1137 | 60.0 | 13200 | 217.4995 | 0.0488 |
113
 
114
 
115
  ### Framework versions
config.json CHANGED
@@ -9,7 +9,7 @@
9
  "architectures": [
10
  "Wav2Vec2ForCTC"
11
  ],
12
- "attention_dropout": 0.15,
13
  "bos_token_id": 1,
14
  "classifier_proj_size": 256,
15
  "codevector_dim": 256,
@@ -50,12 +50,12 @@
50
  "feat_extract_activation": "gelu",
51
  "feat_extract_dropout": 0.0,
52
  "feat_extract_norm": "group",
53
- "feat_proj_dropout": 0.15,
54
  "feat_quantizer_dropout": 0.0,
55
- "final_dropout": 0.15,
56
  "gradient_checkpointing": false,
57
  "hidden_act": "gelu",
58
- "hidden_dropout": 0.15,
59
  "hidden_dropout_prob": 0.1,
60
  "hidden_size": 768,
61
  "initializer_range": 0.02,
 
9
  "architectures": [
10
  "Wav2Vec2ForCTC"
11
  ],
12
+ "attention_dropout": 0.1,
13
  "bos_token_id": 1,
14
  "classifier_proj_size": 256,
15
  "codevector_dim": 256,
 
50
  "feat_extract_activation": "gelu",
51
  "feat_extract_dropout": 0.0,
52
  "feat_extract_norm": "group",
53
+ "feat_proj_dropout": 0.1,
54
  "feat_quantizer_dropout": 0.0,
55
+ "final_dropout": 0.1,
56
  "gradient_checkpointing": false,
57
  "hidden_act": "gelu",
58
+ "hidden_dropout": 0.1,
59
  "hidden_dropout_prob": 0.1,
60
  "hidden_size": 768,
61
  "initializer_range": 0.02,
model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:b8fe0a83b85190e9c5df256386b63f109df3c6d5bc2d19218bec38e8582c02ab
3
  size 377611120
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:95cc569bad2e2e6c1ac4216a9ac78ae92fffa33ab8a87c442f257d3d58466dde
3
  size 377611120
training_args.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:6cb190c3f93a398db49cf968b1143a0a2d6ea9ba39aea61f61608d6d31dfaa5a
3
- size 5176
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c9442254a5e10f72e35377f72189cac35be46ef05f53ff1f21623bdb0926e4b3
3
+ size 5240