SeizureClassifier_Wav2Vec_43243498

This model is a fine-tuned version of facebook/wav2vec2-base on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.0520
• Accuracy: 0.9901

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
• gradient_accumulation_steps: 4
• 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: 15
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
1.1121	0.99	44	0.9842	0.8144
0.6303	1.99	88	0.5874	0.8861
0.4605	2.98	132	0.3826	0.9406
0.323	4.0	177	0.2791	0.9530
0.2435	4.99	221	0.3828	0.8688
0.2354	5.99	265	0.1321	0.9752
0.2491	6.98	309	0.1552	0.9653
0.1116	8.0	354	0.1540	0.9579
0.0934	8.99	398	0.1053	0.9827
0.0774	9.99	442	0.1016	0.9777
0.0553	10.98	486	0.1856	0.9530
0.0368	12.0	531	0.1151	0.9728
0.017	12.99	575	0.0516	0.9876
0.0153	13.99	619	0.0540	0.9901
0.0144	14.92	660	0.0520	0.9901

Framework versions

• Transformers 4.36.2
• Pytorch 2.1.2+cu118
• Datasets 2.16.1
• Tokenizers 0.15.0