swin-tiny-patch4-window7-224

This model is a fine-tuned version of microsoft/swin-tiny-patch4-window7-224 on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.2830
• Accuracy: 0.9667
• Precision: 0.9698
• Sensitivity: 0.9667
• Specificity: 0.9667
• F1: 0.9675
• Auc: 0.9873
• Mcc: 0.8886
• J Stat: 0.9333
• Confusion Matrix: [[145, 5], [1, 29]]

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: 9.131608136923706e-05
• train_batch_size: 16
• eval_batch_size: 16
• seed: 42
• optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.07788173246475068
• num_epochs: 10
• label_smoothing_factor: 0.11328743246488582

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision	Sensitivity	Specificity	F1	Auc	Mcc	J Stat	Confusion Matrix
0.4982	1.0	94	0.3367	0.9234	0.9297	0.9256	0.9227	0.9250	0.9750	0.8135	0.8483	[[1026, 86], [29, 361]]
0.3356	2.0	188	0.2915	0.9654	0.9657	0.8897	0.9919	0.9649	0.9819	0.9090	0.8817	[[1103, 9], [43, 347]]
0.2986	3.0	282	0.2642	0.9767	0.9769	0.9231	0.9955	0.9764	0.9869	0.9390	0.9186	[[1107, 5], [30, 360]]
0.2723	4.0	376	0.2647	0.9827	0.9827	0.9487	0.9946	0.9826	0.9944	0.9547	0.9433	[[1106, 6], [20, 370]]
0.269	5.0	470	0.2514	0.9840	0.9841	0.9487	0.9964	0.9839	0.9957	0.9583	0.9451	[[1108, 4], [20, 370]]
0.2403	6.0	564	0.2434	0.9900	0.9900	0.9718	0.9964	0.9900	0.9972	0.9739	0.9682	[[1108, 4], [11, 379]]
0.2448	7.0	658	0.2284	0.9953	0.9953	0.9872	0.9982	0.9953	0.9993	0.9879	0.9854	[[1110, 2], [5, 385]]
0.2303	8.0	752	0.2257	0.9953	0.9953	0.9872	0.9982	0.9953	0.9999	0.9879	0.9854	[[1110, 2], [5, 385]]
0.2284	9.0	846	0.2215	0.9980	0.9980	0.9974	0.9982	0.9980	1.0000	0.9948	0.9956	[[1110, 2], [1, 389]]
0.2257	10.0	940	0.2204	0.9980	0.9980	0.9974	0.9982	0.9980	1.0000	0.9948	0.9956	[[1110, 2], [1, 389]]

Framework versions

• Transformers 4.52.4
• Pytorch 2.6.0+cu124
• Datasets 4.0.0
• Tokenizers 0.21.2