josedonoso/vit-ecg-khan

vit-base-ecg

This model is a fine-tuned version of google/vit-base-patch16-224-in21k on the imagefolder dataset. It achieves the following results on the evaluation set:

• Loss: 0.1003
• Accuracy: 0.9643

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: 0.0002
• train_batch_size: 16
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 20
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
0.596	2.4390	100	0.5431	0.8214
0.0656	4.8780	200	0.1628	0.95
0.0192	7.3171	300	0.1003	0.9643
0.0926	9.7561	400	0.1262	0.95
0.0064	12.1951	500	0.1611	0.9643
0.0049	14.6341	600	0.1539	0.9643
0.0044	17.0732	700	0.1509	0.9643
0.0041	19.5122	800	0.1499	0.9643

Framework versions

• Transformers 4.41.2
• Pytorch 2.3.0+cu121
• Datasets 2.20.0
• Tokenizers 0.19.1