metadata
license: apache-2.0
base_model: facebook/wav2vec2-base-960h
tags:
- audio-classification
- deepfake
- audio-spoof
- generated_from_trainer
metrics:
- accuracy
model-index:
- name: wav2vec2-base-960h-asv19-deepfake
results: []
wav2vec2-base-960h-asv19-deepfake
This model is a fine-tuned version of facebook/wav2vec2-base-960h on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 0.1594
- Accuracy: 0.9610
- Far: 0.0098
- Frr: 0.0423
- Eer: 0.0261
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: 1e-06
- train_batch_size: 2
- eval_batch_size: 2
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 4
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 2.0
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy | Far | Frr | Eer |
|---|---|---|---|---|---|---|---|
| 0.4231 | 0.39 | 2500 | 0.3092 | 0.8974 | 1.0 | 0.0 | 0.5 |
| 0.2276 | 0.79 | 5000 | 0.0955 | 0.9762 | 0.0530 | 0.0205 | 0.0367 |
| 0.0979 | 1.18 | 7500 | 0.1456 | 0.9555 | 0.0122 | 0.0482 | 0.0302 |
| 0.0784 | 1.58 | 10000 | 0.1613 | 0.9586 | 0.0094 | 0.0451 | 0.0272 |
| 0.0618 | 1.97 | 12500 | 0.1594 | 0.9610 | 0.0098 | 0.0423 | 0.0261 |
Framework versions
- Transformers 4.38.0.dev0
- Pytorch 2.1.2+cu121
- Datasets 2.16.2.dev0
- Tokenizers 0.15.2