<!-- 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. -->
wav2vec2-base-timit-demo-google-colab
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: 0.5602
- Wer: 0.3438
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
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 3.586 | 1.0 | 500 | 2.7824 | 1.0 |
| 1.1036 | 2.01 | 1000 | 0.7074 | 0.5704 |
| 0.4623 | 3.01 | 1500 | 0.4477 | 0.4580 |
| 0.3152 | 4.02 | 2000 | 0.4067 | 0.4235 |
| 0.2596 | 5.02 | 2500 | 0.4183 | 0.4091 |
| 0.2007 | 6.02 | 3000 | 0.4237 | 0.3892 |
| 0.1643 | 7.03 | 3500 | 0.4918 | 0.3898 |
| 0.1436 | 8.03 | 4000 | 0.4023 | 0.3914 |
| 0.121 | 9.04 | 4500 | 0.4674 | 0.3780 |
| 0.112 | 10.04 | 5000 | 0.4919 | 0.3902 |
| 0.0985 | 11.04 | 5500 | 0.5252 | 0.3781 |
| 0.09 | 12.05 | 6000 | 0.5382 | 0.3595 |
| 0.085 | 13.05 | 6500 | 0.5215 | 0.3646 |
| 0.0731 | 14.06 | 7000 | 0.4600 | 0.3673 |
| 0.0645 | 15.06 | 7500 | 0.5151 | 0.3671 |
| 0.0615 | 16.06 | 8000 | 0.5289 | 0.3662 |
| 0.0543 | 17.07 | 8500 | 0.5193 | 0.3752 |
| 0.057 | 18.07 | 9000 | 0.5403 | 0.3734 |
| 0.0466 | 19.08 | 9500 | 0.5607 | 0.3690 |
| 0.0462 | 20.08 | 10000 | 0.5381 | 0.3599 |
| 0.0409 | 21.08 | 10500 | 0.6067 | 0.3681 |
| 0.0373 | 22.09 | 11000 | 0.5607 | 0.3647 |
| 0.0355 | 23.09 | 11500 | 0.5635 | 0.3612 |
| 0.033 | 24.1 | 12000 | 0.5526 | 0.3570 |
| 0.0284 | 25.1 | 12500 | 0.5514 | 0.3484 |
| 0.0272 | 26.1 | 13000 | 0.5378 | 0.3479 |
| 0.0258 | 27.11 | 13500 | 0.5578 | 0.3446 |
| 0.0242 | 28.11 | 14000 | 0.5557 | 0.3442 |
| 0.0235 | 29.12 | 14500 | 0.5602 | 0.3438 |
Framework versions
- Transformers 4.17.0
- Pytorch 1.12.1+cu113
- Datasets 1.18.3
- Tokenizers 0.12.1
