<!-- 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: 0.5442
- Wer: 0.3327
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 |
---|---|---|---|---|
3.558 | 1.0 | 500 | 1.9825 | 0.9952 |
0.8674 | 2.01 | 1000 | 0.5186 | 0.5141 |
0.4291 | 3.01 | 1500 | 0.4576 | 0.4590 |
0.3008 | 4.02 | 2000 | 0.4906 | 0.4449 |
0.2297 | 5.02 | 2500 | 0.4460 | 0.4242 |
0.1848 | 6.02 | 3000 | 0.4410 | 0.4013 |
0.1552 | 7.03 | 3500 | 0.4632 | 0.3833 |
0.1335 | 8.03 | 4000 | 0.4588 | 0.3870 |
0.1209 | 9.04 | 4500 | 0.4553 | 0.3751 |
0.108 | 10.04 | 5000 | 0.4463 | 0.3752 |
0.1011 | 11.04 | 5500 | 0.4730 | 0.3628 |
0.0898 | 12.05 | 6000 | 0.4716 | 0.3739 |
0.0822 | 13.05 | 6500 | 0.5299 | 0.3696 |
0.0702 | 14.06 | 7000 | 0.5478 | 0.3655 |
0.0648 | 15.06 | 7500 | 0.5487 | 0.3631 |
0.0595 | 16.06 | 8000 | 0.6031 | 0.3566 |
0.0567 | 17.07 | 8500 | 0.5375 | 0.3476 |
0.0542 | 18.07 | 9000 | 0.5286 | 0.3540 |
0.0467 | 19.08 | 9500 | 0.5743 | 0.3574 |
0.0419 | 20.08 | 10000 | 0.5855 | 0.3557 |
0.0428 | 21.08 | 10500 | 0.5339 | 0.3459 |
0.0346 | 22.09 | 11000 | 0.5261 | 0.3399 |
0.0312 | 23.09 | 11500 | 0.5699 | 0.3435 |
0.0319 | 24.1 | 12000 | 0.5647 | 0.3442 |
0.0288 | 25.1 | 12500 | 0.5419 | 0.3404 |
0.0247 | 26.1 | 13000 | 0.5388 | 0.3362 |
0.0249 | 27.11 | 13500 | 0.5521 | 0.3357 |
0.0214 | 28.11 | 14000 | 0.5515 | 0.3307 |
0.0235 | 29.12 | 14500 | 0.5442 | 0.3327 |
Framework versions
- Transformers 4.17.0
- Pytorch 1.12.1+cu113
- Datasets 1.18.3
- Tokenizers 0.13.0