Hugging Face's logo

DewiBrynJones
/
wav2vec2-xlsr-53-ft-btb-ccv-cy

Automatic Speech Recognition
Transformers
TensorBoard
Safetensors
wav2vec2
Generated from Trainer
Inference Endpoints
Model card Files Metrics Community
DewiBrynJones commited on
Commit
ce389ad
•
1 Parent(s): 3f2e124

Model save

Browse files
Files changed (1) hide show
  1. README.md +24 -46
README.md CHANGED
@@ -2,8 +2,6 @@
2
  license: apache-2.0
3
  base_model: facebook/wav2vec2-large-xlsr-53
4
  tags:
5
- - automatic-speech-recognition
6
- - DewiBrynJones/banc-trawsgrifiadau-bangor-clean-with-ccv
7
  - generated_from_trainer
8
  metrics:
9
  - wer
@@ -17,10 +15,10 @@ should probably proofread and complete it, then remove this comment. -->
17
 
18
  # wav2vec2-xlsr-53-ft-btb-ccv-cy
19
 
20
- This model is a fine-tuned version of [facebook/wav2vec2-large-xlsr-53](https://huggingface.co/facebook/wav2vec2-large-xlsr-53) on the DEWIBRYNJONES/BANC-TRAWSGRIFIADAU-BANGOR-CLEAN-WITH-CCV - DEFAULT dataset.
21
  It achieves the following results on the evaluation set:
22
- - Loss: nan
23
- - Wer: 1.0
24
 
25
  ## Model description
26
 
@@ -46,53 +44,33 @@ The following hyperparameters were used during training:
46
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
47
  - lr_scheduler_type: linear
48
  - lr_scheduler_warmup_steps: 600
49
- - training_steps: 20000
50
  - mixed_precision_training: Native AMP
51
 
52
  ### Training results
53
 
54
  | Training Loss | Epoch | Step | Validation Loss | Wer |
55
  |:-------------:|:------:|:-----:|:---------------:|:------:|
56
- | 4.7126 | 0.0321 | 500 | 1.7047 | 0.9346 |
57
- | 1.0533 | 0.0641 | 1000 | 1.1487 | 0.7907 |
58
- | 0.8268 | 0.0962 | 1500 | 1.0602 | 0.7815 |
59
- | 0.7188 | 0.1283 | 2000 | 0.9336 | 0.6717 |
60
- | 0.6725 | 0.1603 | 2500 | 0.9304 | 0.6561 |
61
- | 0.6295 | 0.1924 | 3000 | 0.8600 | 0.6257 |
62
- | 0.6003 | 0.2244 | 3500 | 0.8395 | 0.6113 |
63
- | 0.5847 | 0.2565 | 4000 | 0.7884 | 0.5861 |
64
- | 0.5521 | 0.2886 | 4500 | 0.7741 | 0.5687 |
65
- | 0.5477 | 0.3206 | 5000 | 0.7594 | 0.5536 |
66
- | 0.5346 | 0.3527 | 5500 | 0.7482 | 0.5394 |
67
- | 0.5154 | 0.3848 | 6000 | 0.7294 | 0.5352 |
68
- | 0.492 | 0.4168 | 6500 | 0.7248 | 0.5493 |
69
- | 0.4759 | 0.4489 | 7000 | 0.7077 | 0.5134 |
70
- | 0.4655 | 0.4810 | 7500 | 0.6739 | 0.5064 |
71
- | 0.4594 | 0.5130 | 8000 | 0.6575 | 0.5067 |
72
- | 0.4538 | 0.5451 | 8500 | 0.6493 | 0.5003 |
73
- | 0.4739 | 0.5771 | 9000 | 0.7677 | 0.5239 |
74
- | 0.695 | 0.6092 | 9500 | nan | 1.0 |
75
- | 0.0 | 0.6413 | 10000 | nan | 1.0 |
76
- | 0.0 | 0.6733 | 10500 | nan | 1.0 |
77
- | 0.0 | 0.7054 | 11000 | nan | 1.0 |
78
- | 0.0 | 0.7375 | 11500 | nan | 1.0 |
79
- | 0.0 | 0.7695 | 12000 | nan | 1.0 |
80
- | 0.0 | 0.8016 | 12500 | nan | 1.0 |
81
- | 0.0 | 0.8337 | 13000 | nan | 1.0 |
82
- | 0.0 | 0.8657 | 13500 | nan | 1.0 |
83
- | 0.0 | 0.8978 | 14000 | nan | 1.0 |
84
- | 0.0 | 0.9298 | 14500 | nan | 1.0 |
85
- | 0.0 | 0.9619 | 15000 | nan | 1.0 |
86
- | 0.0 | 0.9940 | 15500 | nan | 1.0 |
87
- | 0.0 | 1.0260 | 16000 | nan | 1.0 |
88
- | 0.0 | 1.0581 | 16500 | nan | 1.0 |
89
- | 0.0 | 1.0902 | 17000 | nan | 1.0 |
90
- | 0.0 | 1.1222 | 17500 | nan | 1.0 |
91
- | 0.0 | 1.1543 | 18000 | nan | 1.0 |
92
- | 0.0 | 1.1864 | 18500 | nan | 1.0 |
93
- | 0.0 | 1.2184 | 19000 | nan | 1.0 |
94
- | 0.0 | 1.2505 | 19500 | nan | 1.0 |
95
- | 0.0 | 1.2825 | 20000 | nan | 1.0 |
96
 
97
 
98
  ### Framework versions
 
2
  license: apache-2.0
3
  base_model: facebook/wav2vec2-large-xlsr-53
4
  tags:
 
 
5
  - generated_from_trainer
6
  metrics:
7
  - wer
 
15
 
16
  # wav2vec2-xlsr-53-ft-btb-ccv-cy
17
 
18
+ This model is a fine-tuned version of [facebook/wav2vec2-large-xlsr-53](https://huggingface.co/facebook/wav2vec2-large-xlsr-53) on an unknown dataset.
19
  It achieves the following results on the evaluation set:
20
+ - Loss: 0.5324
21
+ - Wer: 0.4014
22
 
23
  ## Model description
24
 
 
44
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
45
  - lr_scheduler_type: linear
46
  - lr_scheduler_warmup_steps: 600
47
+ - training_steps: 10000
48
  - mixed_precision_training: Native AMP
49
 
50
  ### Training results
51
 
52
  | Training Loss | Epoch | Step | Validation Loss | Wer |
53
  |:-------------:|:------:|:-----:|:---------------:|:------:|
54
+ | 4.7051 | 0.0321 | 500 | 1.7504 | 0.9570 |
55
+ | 1.0409 | 0.0641 | 1000 | 1.1511 | 0.7761 |
56
+ | 0.8183 | 0.0962 | 1500 | 1.0506 | 0.7097 |
57
+ | 0.7091 | 0.1283 | 2000 | 0.9421 | 0.6610 |
58
+ | 0.6547 | 0.1603 | 2500 | 0.8726 | 0.6128 |
59
+ | 0.6088 | 0.1924 | 3000 | 0.8246 | 0.5990 |
60
+ | 0.5781 | 0.2244 | 3500 | 0.8025 | 0.5747 |
61
+ | 0.5429 | 0.2565 | 4000 | 0.7360 | 0.5305 |
62
+ | 0.5104 | 0.2886 | 4500 | 0.7335 | 0.5394 |
63
+ | 0.501 | 0.3206 | 5000 | 0.6933 | 0.5088 |
64
+ | 0.4708 | 0.3527 | 5500 | 0.6770 | 0.5113 |
65
+ | 0.4526 | 0.3848 | 6000 | 0.6609 | 0.4806 |
66
+ | 0.4235 | 0.4168 | 6500 | 0.6373 | 0.4858 |
67
+ | 0.4032 | 0.4489 | 7000 | 0.6048 | 0.4466 |
68
+ | 0.3863 | 0.4810 | 7500 | 0.5946 | 0.4432 |
69
+ | 0.3766 | 0.5130 | 8000 | 0.5737 | 0.4298 |
70
+ | 0.3746 | 0.5451 | 8500 | 0.5668 | 0.4248 |
71
+ | 0.3586 | 0.5771 | 9000 | 0.5485 | 0.4101 |
72
+ | 0.3552 | 0.6092 | 9500 | 0.5378 | 0.4032 |
73
+ | 0.3326 | 0.6413 | 10000 | 0.5324 | 0.4014 |
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
74
 
75
 
76
  ### Framework versions