model_3-D2-NW-Tuned

This model is a fine-tuned version of microsoft/swinv2-tiny-patch4-window16-256 on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.4832
• Accuracy: 0.85
• Precision: 0.8562
• Sensitivity: 0.6
• Specificity: 0.9
• F1: 0.8528
• Auc: 0.8573
• Mcc: 0.4816
• J Stat: 0.5
• Confusion Matrix: [[135, 15], [12, 18]]

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: 4.7873986946753355e-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: 0.04254487066838565
• num_epochs: 7
• label_smoothing_factor: 0.10735647330773991

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision	Sensitivity	Specificity	F1	Auc	Mcc	J Stat	Confusion Matrix
0.6235	1.0	94	0.5001	0.8333	0.6944	0.0	1.0	0.7576	0.7307	0.0	0.0	[[150, 0], [30, 0]]
0.5977	2.0	188	0.4664	0.8278	0.7524	0.0333	0.9867	0.7644	0.7689	0.0582	0.0200	[[148, 2], [29, 1]]
0.5156	3.0	282	0.4218	0.8611	0.8606	0.2	0.9933	0.8229	0.848	0.3727	0.1933	[[149, 1], [24, 6]]
0.4736	4.0	376	0.4577	0.85	0.8699	0.7	0.88	0.8575	0.8616	0.5247	0.5800	[[132, 18], [9, 21]]
0.4352	5.0	470	0.4336	0.8389	0.8142	0.3	0.9467	0.8200	0.8573	0.3143	0.2467	[[142, 8], [21, 9]]
0.3775	6.0	564	0.5115	0.8222	0.8470	0.6333	0.86	0.8319	0.8491	0.4422	0.4933	[[129, 21], [11, 19]]
0.3295	7.0	658	0.4832	0.85	0.8562	0.6	0.9	0.8528	0.8573	0.4816	0.5	[[135, 15], [12, 18]]

Framework versions

• Transformers 4.57.1
• Pytorch 2.6.0+cu124
• Datasets 4.4.1
• Tokenizers 0.22.1