wav2vec2-base-finetuned-eos_poc5_ge-di-v5
This model is a fine-tuned version of facebook/wav2vec2-base on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 0.7709
- Accuracy: 0.7143
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: 3e-05
- train_batch_size: 16
- eval_batch_size: 16
- seed: 42
- optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 10
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|---|---|---|---|---|
| 1.0533 | 1.0 | 302 | 1.0742 | 0.5529 |
| 0.8198 | 2.0 | 604 | 1.0600 | 0.5584 |
| 0.9422 | 3.0 | 906 | 0.9573 | 0.6067 |
| 0.9445 | 4.0 | 1208 | 0.9157 | 0.6252 |
| 0.7605 | 5.0 | 1510 | 0.8589 | 0.6568 |
| 0.7291 | 6.0 | 1812 | 0.8798 | 0.6549 |
| 0.6991 | 7.0 | 2114 | 0.8891 | 0.6568 |
| 0.63 | 8.0 | 2416 | 0.8391 | 0.6957 |
| 0.5887 | 9.0 | 2718 | 0.7898 | 0.7013 |
| 0.5454 | 10.0 | 3020 | 0.7709 | 0.7143 |
Framework versions
- Transformers 4.47.1
- Pytorch 2.5.1+cu121
- Datasets 3.2.0
- Tokenizers 0.21.0