Hugging Face's logo

8688chris
/
Helldivers2ASR_V3

Safetensors
wav2vec2
Generated from Trainer
Model card Files
xet
Community
8688chris commited on
Commit
b6aec3d
·
verified ·
1 Parent(s): 3f3b621

Model save

Browse files
Files changed (4) hide show
  1. README.md +36 -66
  2. config.json +4 -4
  3. model.safetensors +1 -1
  4. training_args.bin +1 -1
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: 528.4360
21
- - Wer: 0.1080
22
 
23
  ## Model description
24
 
@@ -37,79 +37,49 @@ More information needed
37
  ### Training hyperparameters
38
 
39
  The following hyperparameters were used during training:
40
- - learning_rate: 0.0002
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
- | 1470.2922 | 1.0 | 220 | 928.8249 | 0.4508 |
54
- | 948.333 | 2.0 | 440 | 937.5704 | 0.3408 |
55
- | 741.5294 | 3.0 | 660 | 725.8159 | 0.3007 |
56
- | 563.649 | 4.0 | 880 | 675.3207 | 0.2507 |
57
- | 480.7118 | 5.0 | 1100 | 750.5165 | 0.2377 |
58
- | 413.2753 | 6.0 | 1320 | 618.5261 | 0.2105 |
59
- | 357.5316 | 7.0 | 1540 | 560.3909 | 0.1816 |
60
- | 303.942 | 8.0 | 1760 | 490.6678 | 0.1861 |
61
- | 291.6549 | 9.0 | 1980 | 458.5207 | 0.1641 |
62
- | 252.5657 | 10.0 | 2200 | 472.9116 | 0.1495 |
63
- | 224.752 | 11.0 | 2420 | 842.5786 | 0.2153 |
64
- | 208.8816 | 12.0 | 2640 | 475.4429 | 0.1576 |
65
- | 204.4481 | 13.0 | 2860 | 779.1060 | 0.1702 |
66
- | 194.729 | 14.0 | 3080 | 572.8595 | 0.1413 |
67
- | 156.7179 | 15.0 | 3300 | 544.5276 | 0.1407 |
68
- | 175.925 | 16.0 | 3520 | 388.6498 | 0.1236 |
69
- | 143.772 | 17.0 | 3740 | 713.5331 | 0.1714 |
70
- | 141.2694 | 18.0 | 3960 | 390.4991 | 0.1210 |
71
- | 132.3416 | 19.0 | 4180 | 520.8185 | 0.1230 |
72
- | 130.5604 | 20.0 | 4400 | 502.7421 | 0.1202 |
73
- | 123.3909 | 21.0 | 4620 | 329.0610 | 0.0860 |
74
- | 109.1357 | 22.0 | 4840 | 330.5388 | 0.0986 |
75
- | 100.0788 | 23.0 | 5060 | 359.4705 | 0.0952 |
76
- | 100.0983 | 24.0 | 5280 | 529.3177 | 0.1110 |
77
- | 98.9179 | 25.0 | 5500 | 698.3449 | 0.1586 |
78
- | 89.327 | 26.0 | 5720 | 580.2330 | 0.1106 |
79
- | 73.2891 | 27.0 | 5940 | 512.8746 | 0.1171 |
80
- | 81.7329 | 28.0 | 6160 | 385.6871 | 0.0962 |
81
- | 81.4426 | 29.0 | 6380 | 419.5893 | 0.1011 |
82
- | 87.7749 | 30.0 | 6600 | 477.3170 | 0.1318 |
83
- | 78.6555 | 31.0 | 6820 | 350.9287 | 0.0946 |
84
- | 81.6152 | 32.0 | 7040 | 386.9069 | 0.1065 |
85
- | 66.2793 | 33.0 | 7260 | 429.2931 | 0.1080 |
86
- | 62.7566 | 34.0 | 7480 | 519.3239 | 0.1216 |
87
- | 64.9473 | 35.0 | 7700 | 398.3335 | 0.0968 |
88
- | 68.8528 | 36.0 | 7920 | 385.7329 | 0.1114 |
89
- | 54.9865 | 37.0 | 8140 | 465.0021 | 0.1196 |
90
- | 49.088 | 38.0 | 8360 | 441.8585 | 0.1131 |
91
- | 57.9633 | 39.0 | 8580 | 325.4701 | 0.0891 |
92
- | 54.7517 | 40.0 | 8800 | 335.5435 | 0.0878 |
93
- | 45.5077 | 41.0 | 9020 | 520.4970 | 0.1100 |
94
- | 52.4684 | 42.0 | 9240 | 406.8724 | 0.1033 |
95
- | 55.7947 | 43.0 | 9460 | 315.6170 | 0.0842 |
96
- | 53.4148 | 44.0 | 9680 | 430.8446 | 0.1025 |
97
- | 55.6933 | 45.0 | 9900 | 402.8480 | 0.0962 |
98
- | 35.7859 | 46.0 | 10120 | 353.2895 | 0.0897 |
99
- | 41.927 | 47.0 | 10340 | 385.7763 | 0.0905 |
100
- | 47.249 | 48.0 | 10560 | 370.2774 | 0.0907 |
101
- | 49.3789 | 49.0 | 10780 | 395.4215 | 0.0903 |
102
- | 37.4909 | 50.0 | 11000 | 460.0715 | 0.0976 |
103
- | 48.9771 | 51.0 | 11220 | 480.6746 | 0.0994 |
104
- | 36.714 | 52.0 | 11440 | 512.2975 | 0.1041 |
105
- | 41.4792 | 53.0 | 11660 | 503.1389 | 0.1063 |
106
- | 50.2679 | 54.0 | 11880 | 552.9897 | 0.1110 |
107
- | 40.709 | 55.0 | 12100 | 534.4197 | 0.1086 |
108
- | 47.1658 | 56.0 | 12320 | 530.0042 | 0.1088 |
109
- | 40.837 | 57.0 | 12540 | 530.1498 | 0.1080 |
110
- | 42.7641 | 58.0 | 12760 | 528.1923 | 0.1080 |
111
- | 41.2417 | 59.0 | 12980 | 528.1598 | 0.1080 |
112
- | 34.7281 | 60.0 | 13200 | 528.4360 | 0.1080 |
113
 
114
 
115
  ### 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: 134.4454
21
+ - Wer: 0.0694
22
 
23
  ## Model description
24
 
 
37
  ### Training hyperparameters
38
 
39
  The following hyperparameters were used during training:
40
+ - learning_rate: 7e-05
41
+ - train_batch_size: 16
42
+ - eval_batch_size: 16
43
  - seed: 42
44
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
45
  - lr_scheduler_type: cosine
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
+ | 707.3282 | 1.0 | 486 | 401.7251 | 0.3725 |
54
+ | 477.7054 | 2.0 | 972 | 338.9568 | 0.3002 |
55
+ | 377.4668 | 3.0 | 1458 | 332.3391 | 0.2766 |
56
+ | 312.8686 | 4.0 | 1944 | 250.9001 | 0.2083 |
57
+ | 263.9122 | 5.0 | 2430 | 255.3110 | 0.1924 |
58
+ | 225.1422 | 6.0 | 2916 | 235.4016 | 0.1733 |
59
+ | 205.3537 | 7.0 | 3402 | 167.1833 | 0.1290 |
60
+ | 180.8011 | 8.0 | 3888 | 150.0885 | 0.1195 |
61
+ | 159.5088 | 9.0 | 4374 | 171.2871 | 0.1314 |
62
+ | 150.4287 | 10.0 | 4860 | 158.4681 | 0.1093 |
63
+ | 127.9891 | 11.0 | 5346 | 142.8434 | 0.0979 |
64
+ | 120.8978 | 12.0 | 5832 | 134.8509 | 0.0976 |
65
+ | 110.0645 | 13.0 | 6318 | 132.9552 | 0.0976 |
66
+ | 104.547 | 14.0 | 6804 | 122.0037 | 0.0835 |
67
+ | 96.8871 | 15.0 | 7290 | 142.5403 | 0.0873 |
68
+ | 87.4426 | 16.0 | 7776 | 136.5416 | 0.0862 |
69
+ | 80.2423 | 17.0 | 8262 | 113.7286 | 0.0772 |
70
+ | 76.5584 | 18.0 | 8748 | 123.2151 | 0.0747 |
71
+ | 71.483 | 19.0 | 9234 | 123.1056 | 0.0729 |
72
+ | 74.9826 | 20.0 | 9720 | 126.8235 | 0.0710 |
73
+ | 64.5212 | 21.0 | 10206 | 123.9980 | 0.0692 |
74
+ | 66.9201 | 22.0 | 10692 | 133.9775 | 0.0747 |
75
+ | 72.1223 | 23.0 | 11178 | 125.7820 | 0.0677 |
76
+ | 59.429 | 24.0 | 11664 | 137.0608 | 0.0730 |
77
+ | 64.658 | 25.0 | 12150 | 123.1982 | 0.0655 |
78
+ | 62.1091 | 26.0 | 12636 | 128.4541 | 0.0664 |
79
+ | 59.1036 | 27.0 | 13122 | 132.6404 | 0.0681 |
80
+ | 58.5964 | 28.0 | 13608 | 133.0307 | 0.0685 |
81
+ | 58.6317 | 29.0 | 14094 | 134.8861 | 0.0694 |
82
+ | 64.5423 | 30.0 | 14580 | 134.4454 | 0.0694 |
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
83
 
84
 
85
  ### Framework versions
config.json CHANGED
@@ -9,7 +9,7 @@
9
  "architectures": [
10
  "Wav2Vec2ForCTC"
11
  ],
12
- "attention_dropout": 0.1,
13
  "bos_token_id": 1,
14
  "classifier_proj_size": 256,
15
  "codevector_dim": 256,
@@ -50,12 +50,12 @@
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,
 
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
  "feat_extract_activation": "gelu",
51
  "feat_extract_dropout": 0.0,
52
  "feat_extract_norm": "group",
53
+ "feat_proj_dropout": 0.2,
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,
model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:641312a697db1549dc4be559c33c1ef4d71b4b3f0f896883e40c8018b874d228
3
  size 377611120
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:06661a1451962c8e612b690642ea3322b720c1955c06ed1d5c5d4b572ae59414
3
  size 377611120
training_args.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:5217d6cb05814b4546e4a47ad97665a13f49f8b8a23df9969e97236cd7502078
3
  size 5240
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:088a123df7ffb58042710ea939a332a151f6ba323aed87a2ac5c603e14785fb3
3
  size 5240