<!-- 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: 0.5282
- Wer: 0.3302
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 |
|---|---|---|---|---|
| 3.5185 | 1.0 | 500 | 1.5798 | 0.9593 |
| 0.8096 | 2.01 | 1000 | 0.5024 | 0.5082 |
| 0.4196 | 3.01 | 1500 | 0.4594 | 0.4489 |
| 0.2936 | 4.02 | 2000 | 0.4104 | 0.4131 |
| 0.2215 | 5.02 | 2500 | 0.4308 | 0.4062 |
| 0.1891 | 6.02 | 3000 | 0.4242 | 0.3825 |
| 0.1626 | 7.03 | 3500 | 0.4187 | 0.3792 |
| 0.136 | 8.03 | 4000 | 0.4387 | 0.3766 |
| 0.1221 | 9.04 | 4500 | 0.4634 | 0.3832 |
| 0.1119 | 10.04 | 5000 | 0.4271 | 0.3640 |
| 0.0976 | 11.04 | 5500 | 0.4379 | 0.3701 |
| 0.0846 | 12.05 | 6000 | 0.4686 | 0.3648 |
| 0.0792 | 13.05 | 6500 | 0.4502 | 0.3595 |
| 0.0709 | 14.06 | 7000 | 0.4723 | 0.3634 |
| 0.0671 | 15.06 | 7500 | 0.4601 | 0.3577 |
| 0.058 | 16.06 | 8000 | 0.5146 | 0.3535 |
| 0.055 | 17.07 | 8500 | 0.5352 | 0.3540 |
| 0.0576 | 18.07 | 9000 | 0.5102 | 0.3469 |
| 0.0448 | 19.08 | 9500 | 0.5159 | 0.3527 |
| 0.0429 | 20.08 | 10000 | 0.5085 | 0.3538 |
| 0.0384 | 21.08 | 10500 | 0.5001 | 0.3453 |
| 0.0339 | 22.09 | 11000 | 0.5322 | 0.3460 |
| 0.032 | 23.09 | 11500 | 0.5295 | 0.3459 |
| 0.0306 | 24.1 | 12000 | 0.5285 | 0.3434 |
| 0.0268 | 25.1 | 12500 | 0.5280 | 0.3382 |
| 0.0231 | 26.1 | 13000 | 0.5259 | 0.3363 |
| 0.0242 | 27.11 | 13500 | 0.5298 | 0.3325 |
| 0.0215 | 28.11 | 14000 | 0.5350 | 0.3306 |
| 0.0226 | 29.12 | 14500 | 0.5282 | 0.3302 |
Framework versions
- Transformers 4.17.0
- Pytorch 1.12.1+cu113
- Datasets 1.18.3
- Tokenizers 0.12.1
