SeizureClassifier_Wav2Vec_43243531

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: 0.0063
• Accuracy: 0.9990

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
0.13	1.0	339	0.1382	0.9616
0.0931	2.0	678	0.0613	0.9839
0.0265	3.0	1017	0.0248	0.9942
0.026	4.0	1357	0.0612	0.9900
0.0252	5.0	1696	0.0460	0.9894
0.0369	6.0	2035	0.0148	0.9968
0.0018	7.0	2374	0.0049	0.9990
0.0007	8.0	2714	0.0114	0.9981
0.0056	9.0	3053	0.0107	0.9987
0.0003	10.0	3392	0.0067	0.9990
0.0101	11.0	3731	0.0039	0.9994
0.0073	12.0	4071	0.0049	0.9994
0.0113	13.0	4410	0.0061	0.9990
0.0002	14.0	4749	0.0067	0.9990
0.0002	14.99	5085	0.0063	0.9990

Framework versions

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