DylanonWic commited on
Commit
ad24284
1 Parent(s): 3a2f4d6

Upload 13 files

Browse files
README.md CHANGED
@@ -1,8 +1,6 @@
1
  ---
2
  tags:
3
  - generated_from_trainer
4
- metrics:
5
- - wer
6
  model-index:
7
  - name: wav2vec2-large-asr-th-2
8
  results: []
@@ -15,9 +13,14 @@ should probably proofread and complete it, then remove this comment. -->
15
 
16
  This model was trained from scratch on the None dataset.
17
  It achieves the following results on the evaluation set:
18
- - Loss: 0.2590
19
- - Wer: 0.4210
20
- - Cer: 0.1146
 
 
 
 
 
21
 
22
  ## Model description
23
 
@@ -36,7 +39,7 @@ More information needed
36
  ### Training hyperparameters
37
 
38
  The following hyperparameters were used during training:
39
- - learning_rate: 0.0002
40
  - train_batch_size: 16
41
  - eval_batch_size: 8
42
  - seed: 42
@@ -44,23 +47,13 @@ The following hyperparameters were used during training:
44
  - total_train_batch_size: 48
45
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
46
  - lr_scheduler_type: linear
47
- - lr_scheduler_warmup_steps: 600
48
- - training_steps: 4000
49
  - mixed_precision_training: Native AMP
50
 
51
- ### Training results
52
-
53
- | Training Loss | Epoch | Step | Validation Loss | Wer | Cer |
54
- |:-------------:|:-----:|:----:|:---------------:|:------:|:------:|
55
- | 0.6189 | 0.24 | 1000 | 0.3571 | 0.5060 | 0.1449 |
56
- | 0.5435 | 0.47 | 2000 | 0.3255 | 0.4712 | 0.1328 |
57
- | 0.5612 | 0.71 | 3000 | 0.2870 | 0.4464 | 0.1233 |
58
- | 0.6737 | 0.95 | 4000 | 0.2590 | 0.4210 | 0.1146 |
59
-
60
-
61
  ### Framework versions
62
 
63
- - Transformers 4.27.3
64
  - Pytorch 1.13.1+cu116
65
- - Datasets 2.10.1
66
  - Tokenizers 0.13.2
 
1
  ---
2
  tags:
3
  - generated_from_trainer
 
 
4
  model-index:
5
  - name: wav2vec2-large-asr-th-2
6
  results: []
 
13
 
14
  This model was trained from scratch on the None dataset.
15
  It achieves the following results on the evaluation set:
16
+ - eval_loss: 0.1819
17
+ - eval_wer: 0.3346
18
+ - eval_cer: 0.0910
19
+ - eval_runtime: 735.0436
20
+ - eval_samples_per_second: 13.822
21
+ - eval_steps_per_second: 1.728
22
+ - epoch: 0.95
23
+ - step: 4000
24
 
25
  ## Model description
26
 
 
39
  ### Training hyperparameters
40
 
41
  The following hyperparameters were used during training:
42
+ - learning_rate: 0.0001
43
  - train_batch_size: 16
44
  - eval_batch_size: 8
45
  - seed: 42
 
47
  - total_train_batch_size: 48
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: 5000
52
  - mixed_precision_training: Native AMP
53
 
 
 
 
 
 
 
 
 
 
 
54
  ### Framework versions
55
 
56
+ - Transformers 4.27.4
57
  - Pytorch 1.13.1+cu116
58
+ - Datasets 2.11.0
59
  - Tokenizers 0.13.2
config.json CHANGED
@@ -109,7 +109,7 @@
109
  1
110
  ],
111
  "torch_dtype": "float32",
112
- "transformers_version": "4.27.3",
113
  "use_weighted_layer_sum": false,
114
  "vocab_size": 72,
115
  "xvector_output_dim": 512
 
109
  1
110
  ],
111
  "torch_dtype": "float32",
112
+ "transformers_version": "4.27.4",
113
  "use_weighted_layer_sum": false,
114
  "vocab_size": 72,
115
  "xvector_output_dim": 512
language_model/attrs.json CHANGED
@@ -1 +1 @@
1
- {"alpha": 0.5, "beta": 1.5, "unk_score_offset": -10.0, "score_boundary": true}
 
1
+ {"alpha": 0.7, "beta": 1.0, "unk_score_offset": -10.0, "score_boundary": true}
language_model/internet-crawled.4gram.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0a3a50973877491ce00617514e77aafb859b9ae0e28ae6902e16da3cde5d0787
3
+ size 3499191609
pytorch_model.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:f3db9d140ed973d933448e83b4a3d7c72b7683b7a077f52c258594763abf2139
3
  size 1262197037
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:330f08323ba2112b309cdecc88208920a6afdd21d8d4b8f0b1ea27fabd5868d3
3
  size 1262197037
training_args.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:ab1acd7b1adc89516ec5569b2e31e361d047913189de035bffe974d3c7e42c8a
3
  size 3643
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e7c0bd0bd8e70f57d6247470affb116af5391526153c4774fbf17cfec59c3770
3
  size 3643