<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. -->
data-augmentation-whitenoise-timit-2310
This model is a fine-tuned version of facebook/wav2vec2-base on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.5916
- Wer: 0.3408
Model description
More information needed
Intended uses & limitations
More information needed
Training and evaluation data
More information needed
Training procedure
Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 0.0001
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 1000
- num_epochs: 30
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 3.6731 | 0.67 | 500 | 2.7553 | 1.0 |
| 1.0656 | 1.34 | 1000 | 0.5963 | 0.5297 |
| 0.5065 | 2.01 | 1500 | 0.4898 | 0.4654 |
| 0.3212 | 2.68 | 2000 | 0.4265 | 0.4331 |
| 0.2492 | 3.35 | 2500 | 0.4020 | 0.4073 |
| 0.2116 | 4.02 | 3000 | 0.4152 | 0.3935 |
| 0.1719 | 4.69 | 3500 | 0.4258 | 0.3858 |
| 0.1544 | 5.36 | 4000 | 0.4542 | 0.3818 |
| 0.1474 | 6.03 | 4500 | 0.4612 | 0.3821 |
| 0.1248 | 6.7 | 5000 | 0.4813 | 0.3749 |
| 0.1148 | 7.37 | 5500 | 0.5131 | 0.3772 |
| 0.1145 | 8.04 | 6000 | 0.5383 | 0.3714 |
| 0.0986 | 8.71 | 6500 | 0.5288 | 0.3777 |
| 0.091 | 9.38 | 7000 | 0.5071 | 0.3869 |
| 0.0789 | 10.05 | 7500 | 0.5256 | 0.3819 |
| 0.0747 | 10.72 | 8000 | 0.5287 | 0.3711 |
| 0.0687 | 11.39 | 8500 | 0.5179 | 0.3754 |
| 0.072 | 12.06 | 9000 | 0.7438 | 0.3702 |
| 0.0646 | 12.73 | 9500 | 0.5293 | 0.3777 |
| 0.0621 | 13.4 | 10000 | 0.5536 | 0.3692 |
| 0.0587 | 14.08 | 10500 | 0.5214 | 0.3712 |
| 0.0538 | 14.75 | 11000 | 0.4853 | 0.3694 |
| 0.0614 | 15.42 | 11500 | 0.5439 | 0.3637 |
| 0.0493 | 16.09 | 12000 | 0.5087 | 0.3649 |
| 0.0441 | 16.76 | 12500 | 0.5736 | 0.3621 |
| 0.038 | 17.43 | 13000 | 0.7295 | 0.3650 |
| 0.0397 | 18.1 | 13500 | 0.5722 | 0.3586 |
| 0.0357 | 18.77 | 14000 | 0.5701 | 0.3616 |
| 0.0349 | 19.44 | 14500 | 0.5661 | 0.3599 |
| 0.0318 | 20.11 | 15000 | 0.5346 | 0.3572 |
| 0.0288 | 20.78 | 15500 | 0.6972 | 0.3597 |
| 0.0331 | 21.45 | 16000 | 0.5288 | 0.3576 |
| 0.0304 | 22.12 | 16500 | 0.5813 | 0.3551 |
| 0.0268 | 22.79 | 17000 | 0.5439 | 0.3557 |
| 0.0255 | 23.46 | 17500 | 0.5790 | 0.3531 |
| 0.0244 | 24.13 | 18000 | 0.5794 | 0.3493 |
| 0.0335 | 24.8 | 18500 | 0.5943 | 0.3515 |
| 0.026 | 25.47 | 19000 | 0.5737 | 0.3462 |
| 0.0199 | 26.14 | 19500 | 0.5794 | 0.3469 |
| 0.0213 | 26.81 | 20000 | 0.5955 | 0.3448 |
| 0.0199 | 27.48 | 20500 | 0.5927 | 0.3407 |
| 0.0143 | 28.15 | 21000 | 0.5975 | 0.3415 |
| 0.0167 | 28.82 | 21500 | 0.5835 | 0.3411 |
| 0.0141 | 29.49 | 22000 | 0.5916 | 0.3408 |
Framework versions
- Transformers 4.17.0
- Pytorch 1.12.0+cu113
- Datasets 1.18.3
- Tokenizers 0.12.1
