Hugging Face's logo

tz579
/
wav2vec2-base-timit-fine-tuned

Automatic Speech Recognition
Transformers
TensorBoard
Safetensors
wav2vec2
timit_asr
Generated from Trainer
Eval Results
Inference Endpoints
Model card Files Metrics Community
tz579 commited on
Commit
d2240f4
•
1 Parent(s): 039c272

Model save

Browse files
Files changed (1) hide show
  1. README.md +20 -34
README.md CHANGED
@@ -2,8 +2,6 @@
2
  license: apache-2.0
3
  base_model: facebook/wav2vec2-base
4
  tags:
5
- - automatic-speech-recognition
6
- - timit_asr
7
  - generated_from_trainer
8
  datasets:
9
  - timit_asr
@@ -16,15 +14,15 @@ model-index:
16
  name: Automatic Speech Recognition
17
  type: automatic-speech-recognition
18
  dataset:
19
- name: TIMIT_ASR - NA
20
  type: timit_asr
21
  config: clean
22
  split: test
23
- args: 'Config: na, Training split: train, Eval split: test'
24
  metrics:
25
  - name: Wer
26
  type: wer
27
- value: 0.4090867704634435
28
  ---
29
 
30
  <!-- This model card has been generated automatically according to the information the Trainer had access to. You
@@ -32,10 +30,10 @@ should probably proofread and complete it, then remove this comment. -->
32
 
33
  # wav2vec2-base-timit-fine-tuned
34
 
35
- This model is a fine-tuned version of [facebook/wav2vec2-base](https://huggingface.co/facebook/wav2vec2-base) on the TIMIT_ASR - NA dataset.
36
  It achieves the following results on the evaluation set:
37
- - Loss: 0.4218
38
- - Wer: 0.4091
39
 
40
  ## Model description
41
 
@@ -55,8 +53,8 @@ More information needed
55
 
56
  The following hyperparameters were used during training:
57
  - learning_rate: 0.0001
58
- - train_batch_size: 32
59
- - eval_batch_size: 1
60
  - seed: 42
61
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
62
  - lr_scheduler_type: linear
@@ -68,34 +66,22 @@ The following hyperparameters were used during training:
68
 
69
  | Training Loss | Epoch | Step | Validation Loss | Wer |
70
  |:-------------:|:-------:|:----:|:---------------:|:------:|
71
- | 3.1612 | 0.8621 | 100 | 3.1181 | 1.0 |
72
- | 2.978 | 1.7241 | 200 | 2.9722 | 1.0 |
73
- | 2.9185 | 2.5862 | 300 | 2.9098 | 1.0 |
74
- | 2.1282 | 3.4483 | 400 | 2.0066 | 1.0247 |
75
- | 1.1234 | 4.3103 | 500 | 1.0197 | 0.8393 |
76
- | 0.602 | 5.1724 | 600 | 0.6714 | 0.6600 |
77
- | 0.5032 | 6.0345 | 700 | 0.5285 | 0.5659 |
78
- | 0.3101 | 6.8966 | 800 | 0.4819 | 0.5282 |
79
- | 0.3432 | 7.7586 | 900 | 0.4653 | 0.5272 |
80
- | 0.1922 | 8.6207 | 1000 | 0.4672 | 0.4918 |
81
- | 0.2284 | 9.4828 | 1100 | 0.4834 | 0.4870 |
82
- | 0.1372 | 10.3448 | 1200 | 0.4380 | 0.4727 |
83
- | 0.1105 | 11.2069 | 1300 | 0.4509 | 0.4594 |
84
- | 0.0992 | 12.0690 | 1400 | 0.4196 | 0.4544 |
85
- | 0.1226 | 12.9310 | 1500 | 0.4237 | 0.4321 |
86
- | 0.1013 | 13.7931 | 1600 | 0.4113 | 0.4298 |
87
- | 0.0661 | 14.6552 | 1700 | 0.4038 | 0.4276 |
88
- | 0.0901 | 15.5172 | 1800 | 0.4321 | 0.4225 |
89
- | 0.053 | 16.3793 | 1900 | 0.4076 | 0.4236 |
90
- | 0.0805 | 17.2414 | 2000 | 0.4336 | 0.4156 |
91
- | 0.049 | 18.1034 | 2100 | 0.4193 | 0.4114 |
92
- | 0.0717 | 18.9655 | 2200 | 0.4139 | 0.4091 |
93
- | 0.0389 | 19.8276 | 2300 | 0.4216 | 0.4087 |
94
 
95
 
96
  ### Framework versions
97
 
98
  - Transformers 4.42.0.dev0
99
- - Pytorch 2.3.0a0+git71dd2de
100
  - Datasets 2.19.1
101
  - Tokenizers 0.19.1
 
2
  license: apache-2.0
3
  base_model: facebook/wav2vec2-base
4
  tags:
 
 
5
  - generated_from_trainer
6
  datasets:
7
  - timit_asr
 
14
  name: Automatic Speech Recognition
15
  type: automatic-speech-recognition
16
  dataset:
17
+ name: timit_asr
18
  type: timit_asr
19
  config: clean
20
  split: test
21
+ args: clean
22
  metrics:
23
  - name: Wer
24
  type: wer
25
+ value: 0.4359464627151052
26
  ---
27
 
28
  <!-- This model card has been generated automatically according to the information the Trainer had access to. You
 
30
 
31
  # wav2vec2-base-timit-fine-tuned
32
 
33
+ This model is a fine-tuned version of [facebook/wav2vec2-base](https://huggingface.co/facebook/wav2vec2-base) on the timit_asr dataset.
34
  It achieves the following results on the evaluation set:
35
+ - Loss: 0.4279
36
+ - Wer: 0.4359
37
 
38
  ## Model description
39
 
 
53
 
54
  The following hyperparameters were used during training:
55
  - learning_rate: 0.0001
56
+ - train_batch_size: 64
57
+ - eval_batch_size: 32
58
  - seed: 42
59
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
60
  - lr_scheduler_type: linear
 
66
 
67
  | Training Loss | Epoch | Step | Validation Loss | Wer |
68
  |:-------------:|:-------:|:----:|:---------------:|:------:|
69
+ | 3.158 | 1.7241 | 100 | 3.6803 | 1.0 |
70
+ | 2.9744 | 3.4483 | 200 | 3.1165 | 1.0 |
71
+ | 2.9266 | 5.1724 | 300 | 3.0175 | 1.0 |
72
+ | 2.1336 | 6.8966 | 400 | 2.2135 | 1.0117 |
73
+ | 1.0119 | 8.6207 | 500 | 1.0227 | 0.8251 |
74
+ | 0.4995 | 10.3448 | 600 | 0.7700 | 0.6574 |
75
+ | 0.3233 | 12.0690 | 700 | 0.4970 | 0.5241 |
76
+ | 0.2452 | 13.7931 | 800 | 0.4585 | 0.4908 |
77
+ | 0.181 | 15.5172 | 900 | 0.4626 | 0.4814 |
78
+ | 0.1419 | 17.2414 | 1000 | 0.4917 | 0.4775 |
79
+ | 0.1175 | 18.9655 | 1100 | 0.4279 | 0.4359 |
 
 
 
 
 
 
 
 
 
 
 
 
80
 
81
 
82
  ### Framework versions
83
 
84
  - Transformers 4.42.0.dev0
85
+ - Pytorch 2.3.0a0+gitcd033a1
86
  - Datasets 2.19.1
87
  - Tokenizers 0.19.1