Hugging Face's logo

bookbot
/
wav2vec2-ljspeech-gruut

Automatic Speech Recognition
Transformers
PyTorch
TensorBoard
Safetensors
English
wav2vec2
phoneme-recognition
Generated from Trainer
Eval Results
Inference Endpoints
Model card Files Metrics Community
3
w11wo commited on
Commit
20872d5
·
1 Parent(s): e6203ff

update model card README.md

Browse files
Files changed (1) hide show
  1. README.md +93 -0
README.md ADDED
@@ -0,0 +1,93 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: apache-2.0
3
+ tags:
4
+ - generated_from_trainer
5
+ metrics:
6
+ - wer
7
+ model-index:
8
+ - name: wav2vec2-ljspeech-gruut
9
+ results: []
10
+ ---
11
+
12
+ <!-- This model card has been generated automatically according to the information the Trainer had access to. You
13
+ should probably proofread and complete it, then remove this comment. -->
14
+
15
+ # wav2vec2-ljspeech-gruut
16
+
17
+ This model is a fine-tuned version of [facebook/wav2vec2-base](https://huggingface.co/facebook/wav2vec2-base) on the None dataset.
18
+ It achieves the following results on the evaluation set:
19
+ - Loss: 0.0683
20
+ - Wer: 0.0099
21
+ - Cer: 0.0058
22
+
23
+ ## Model description
24
+
25
+ More information needed
26
+
27
+ ## Intended uses & limitations
28
+
29
+ More information needed
30
+
31
+ ## Training and evaluation data
32
+
33
+ More information needed
34
+
35
+ ## Training procedure
36
+
37
+ ### Training hyperparameters
38
+
39
+ The following hyperparameters were used during training:
40
+ - learning_rate: 0.0001
41
+ - train_batch_size: 16
42
+ - eval_batch_size: 8
43
+ - seed: 42
44
+ - gradient_accumulation_steps: 2
45
+ - total_train_batch_size: 32
46
+ - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
47
+ - lr_scheduler_type: linear
48
+ - lr_scheduler_warmup_steps: 1000
49
+ - num_epochs: 30.0
50
+ - mixed_precision_training: Native AMP
51
+
52
+ ### Training results
53
+
54
+ | Training Loss | Epoch | Step | Validation Loss | Wer | Cer |
55
+ |:-------------:|:-----:|:-----:|:---------------:|:------:|:------:|
56
+ | No log | 1.0 | 348 | 2.2818 | 1.0 | 1.0 |
57
+ | 2.6692 | 2.0 | 696 | 0.2045 | 0.0527 | 0.0299 |
58
+ | 0.2225 | 3.0 | 1044 | 0.1162 | 0.0319 | 0.0189 |
59
+ | 0.2225 | 4.0 | 1392 | 0.0927 | 0.0235 | 0.0147 |
60
+ | 0.0868 | 5.0 | 1740 | 0.0797 | 0.0218 | 0.0143 |
61
+ | 0.0598 | 6.0 | 2088 | 0.0715 | 0.0197 | 0.0128 |
62
+ | 0.0598 | 7.0 | 2436 | 0.0652 | 0.0160 | 0.0103 |
63
+ | 0.0447 | 8.0 | 2784 | 0.0571 | 0.0152 | 0.0095 |
64
+ | 0.0368 | 9.0 | 3132 | 0.0608 | 0.0163 | 0.0112 |
65
+ | 0.0368 | 10.0 | 3480 | 0.0586 | 0.0137 | 0.0083 |
66
+ | 0.0303 | 11.0 | 3828 | 0.0641 | 0.0141 | 0.0085 |
67
+ | 0.0273 | 12.0 | 4176 | 0.0656 | 0.0131 | 0.0079 |
68
+ | 0.0232 | 13.0 | 4524 | 0.0690 | 0.0133 | 0.0082 |
69
+ | 0.0232 | 14.0 | 4872 | 0.0598 | 0.0128 | 0.0079 |
70
+ | 0.0189 | 15.0 | 5220 | 0.0671 | 0.0121 | 0.0074 |
71
+ | 0.017 | 16.0 | 5568 | 0.0654 | 0.0114 | 0.0069 |
72
+ | 0.017 | 17.0 | 5916 | 0.0751 | 0.0118 | 0.0073 |
73
+ | 0.0146 | 18.0 | 6264 | 0.0653 | 0.0112 | 0.0068 |
74
+ | 0.0127 | 19.0 | 6612 | 0.0682 | 0.0112 | 0.0069 |
75
+ | 0.0127 | 20.0 | 6960 | 0.0678 | 0.0114 | 0.0068 |
76
+ | 0.0114 | 21.0 | 7308 | 0.0656 | 0.0111 | 0.0066 |
77
+ | 0.0101 | 22.0 | 7656 | 0.0669 | 0.0109 | 0.0066 |
78
+ | 0.0092 | 23.0 | 8004 | 0.0677 | 0.0108 | 0.0065 |
79
+ | 0.0092 | 24.0 | 8352 | 0.0653 | 0.0104 | 0.0063 |
80
+ | 0.0088 | 25.0 | 8700 | 0.0673 | 0.0102 | 0.0063 |
81
+ | 0.0074 | 26.0 | 9048 | 0.0669 | 0.0105 | 0.0064 |
82
+ | 0.0074 | 27.0 | 9396 | 0.0707 | 0.0101 | 0.0061 |
83
+ | 0.0066 | 28.0 | 9744 | 0.0673 | 0.0100 | 0.0060 |
84
+ | 0.0058 | 29.0 | 10092 | 0.0689 | 0.0100 | 0.0059 |
85
+ | 0.0058 | 30.0 | 10440 | 0.0683 | 0.0099 | 0.0058 |
86
+
87
+
88
+ ### Framework versions
89
+
90
+ - Transformers 4.26.0.dev0
91
+ - Pytorch 1.10.0
92
+ - Datasets 2.7.1
93
+ - Tokenizers 0.13.2