SeizureClassifier_Wav2Vec_B_43828665

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: 0.0355
• Accuracy: 0.9950

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
1.1208	0.99	44	0.9389	0.8441
0.6807	1.99	88	0.5630	0.8911
0.3684	2.98	132	0.3547	0.9332
0.2786	4.0	177	0.2168	0.9678
0.1849	4.99	221	0.2235	0.9530
0.1888	5.99	265	0.1294	0.9802
0.1201	6.98	309	0.1461	0.9703
0.1017	8.0	354	0.1188	0.9777
0.0972	8.99	398	0.1194	0.9752
0.0819	9.99	442	0.0872	0.9851
0.0518	10.98	486	0.0550	0.9851
0.0604	12.0	531	0.0327	0.9975
0.0267	12.99	575	0.0542	0.9926
0.019	13.99	619	0.0354	0.9926
0.0167	14.92	660	0.0355	0.9950

Framework versions

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