deep_fake_model_aug2

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: 0.1652
• Accuracy: 0.9643
• F1-score: 0.9643
• Recall-score: 0.9643
• Precision-score: 0.9644

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: 16
• total_train_batch_size: 64
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• 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	F1-score	Recall-score	Precision-score
0.5877	0.9984	156	0.4429	0.7971	0.7941	0.7971	0.8103
0.3294	1.9968	312	0.2590	0.8980	0.8980	0.8980	0.8984
0.1843	2.9952	468	0.3093	0.8930	0.8922	0.8930	0.9007
0.1505	4.0	625	0.1762	0.9418	0.9418	0.9418	0.9418
0.1021	4.9984	781	0.1905	0.9447	0.9447	0.9447	0.9455
0.0717	5.9968	937	0.1757	0.9523	0.9523	0.9523	0.9537
0.0597	6.9952	1093	0.1500	0.9632	0.9632	0.9632	0.9635
0.0447	8.0	1250	0.1751	0.9617	0.9617	0.9617	0.9617
0.0358	8.9984	1406	0.1652	0.9643	0.9643	0.9643	0.9644
0.0241	9.984	1560	0.1974	0.9602	0.9602	0.9602	0.9610

Framework versions

• Transformers 4.42.3
• Pytorch 2.1.2
• Datasets 2.20.0
• Tokenizers 0.19.1