deepfake-audio-detector-v4
This model is a fine-tuned version of Heem2/Deepfake-audio-detection on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.3732
- Accuracy: 0.9755
- Precision: 0.9894
- Recall: 0.9613
- F1: 0.9752
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: 4
- eval_batch_size: 4
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 8
- optimizer: Use OptimizerNames.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_steps: 100
- num_epochs: 5
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy | Precision | Recall | F1 |
|---|---|---|---|---|---|---|---|
| 0.0001 | 1.0 | 388 | 0.5744 | 0.9381 | 0.9749 | 0.8995 | 0.9357 |
| 1.1624 | 2.0 | 776 | 0.4446 | 0.9678 | 0.9764 | 0.9588 | 0.9675 |
| 0.4063 | 3.0 | 1164 | 0.6542 | 0.9523 | 0.9889 | 0.9149 | 0.9505 |
| 0.0 | 4.0 | 1552 | 0.2746 | 0.9832 | 0.9845 | 0.9820 | 0.9832 |
| 0.1846 | 4.9884 | 1935 | 0.3732 | 0.9755 | 0.9894 | 0.9613 | 0.9752 |
Framework versions
- Transformers 4.51.3
- Pytorch 2.6.0+cu124
- Datasets 3.6.0
- Tokenizers 0.21.1
- Downloads last month
- 6