metadata
license: apache-2.0
base_model: facebook/hubert-base-ls960
tags:
- audio-classification
- deepfake
- audio-spoof
- generated_from_trainer
metrics:
- accuracy
model-index:
- name: hubert-base-960h-asv19-deepfake
results: []
hubert-base-960h-asv19-deepfake
This model is a fine-tuned version of facebook/hubert-base-ls960 on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 0.0624
- Accuracy: 0.9918
- Far: 0.0024
- Frr: 0.0088
- Eer: 0.0056
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: 4.0
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy | Far | Frr | Eer |
|---|---|---|---|---|---|---|---|
| 0.0726 | 0.79 | 5000 | 0.1522 | 0.9648 | 0.0027 | 0.0389 | 0.0208 |
| 0.0338 | 1.58 | 10000 | 0.0995 | 0.9852 | 0.0024 | 0.0162 | 0.0093 |
| 0.0256 | 2.36 | 15000 | 0.0434 | 0.9940 | 0.0027 | 0.0063 | 0.0045 |
| 0.0294 | 3.15 | 20000 | 0.0756 | 0.9899 | 0.0020 | 0.0111 | 0.0065 |
| 0.0175 | 3.94 | 25000 | 0.0624 | 0.9918 | 0.0024 | 0.0088 | 0.0056 |
Framework versions
- Transformers 4.38.0.dev0
- Pytorch 2.1.2+cu121
- Datasets 2.16.2.dev0
- Tokenizers 0.15.2