skin-lesion-vit-finetuned

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

• Loss: 0.1761
• Accuracy: 0.9458
• F1 Macro: 0.9465
• Melanoma Recall: 0.9607

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: 5e-05
• train_batch_size: 16
• eval_batch_size: 32
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 32
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 8
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1 Macro	Melanoma Recall
0.7039	0.9983	299	0.7625	0.7335	0.6859	0.6
0.7119	2.0	599	0.4936	0.8327	0.7772	0.7536
0.4416	2.9983	898	0.4881	0.8287	0.8242	0.8036
0.3581	4.0	1198	0.4164	0.8616	0.8286	0.8929
0.2949	4.9983	1497	0.3000	0.9021	0.8916	0.8893
0.2687	6.0	1797	0.2685	0.9041	0.9206	0.95
0.1699	6.9983	2096	0.2058	0.9318	0.9346	0.9643
0.0908	7.9866	2392	0.1761	0.9458	0.9465	0.9607

Framework versions

• Transformers 4.44.2
• Pytorch 2.11.0+cu128
• Datasets 2.21.0
• Tokenizers 0.19.1