seekwell_skincancer_v2

This model is a fine-tuned version of Anwarkh1/Skin_Cancer-Image_Classification on the None dataset. It achieves the following results on the evaluation set:

• Loss: 1.1158
• Accuracy: 0.7196
• Balanced Accuracy: 0.7106
• F1: 0.7212
• Precision: 0.7437
• Recall: 0.7196

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: 2e-05
• train_batch_size: 64
• 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: cosine
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 15
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Balanced Accuracy	F1	Precision	Recall
1.7574	1.0	29	1.7099	0.4609	0.3919	0.4185	0.4585	0.4609
1.5901	2.0	58	1.4869	0.6043	0.5481	0.6129	0.6335	0.6043
1.332	3.0	87	1.2812	0.6435	0.6394	0.6616	0.7180	0.6435
1.0963	4.0	116	1.1158	0.7196	0.7106	0.7212	0.7437	0.7196
0.8832	5.0	145	0.9947	0.6674	0.7056	0.6867	0.7550	0.6674
0.7099	6.0	174	0.9337	0.6783	0.6794	0.6952	0.7558	0.6783
0.5697	7.0	203	0.8795	0.7261	0.6821	0.7343	0.7603	0.7261
0.4541	8.0	232	0.8924	0.7174	0.6558	0.7209	0.7394	0.7174
0.3736	9.0	261	0.8995	0.7326	0.6931	0.7346	0.7402	0.7326
0.317	10.0	290	0.8938	0.7152	0.6368	0.7198	0.7372	0.7152
0.2772	11.0	319	0.9289	0.7283	0.6432	0.7283	0.7360	0.7283
0.2488	12.0	348	0.9261	0.7217	0.6429	0.7232	0.7294	0.7217
0.232	13.0	377	0.9319	0.7109	0.6357	0.7132	0.7218	0.7109
0.225	14.0	406	0.9435	0.7130	0.6313	0.7144	0.7235	0.7130
0.2228	15.0	435	0.9440	0.7109	0.6278	0.7122	0.7205	0.7109

Framework versions

• Transformers 4.54.1
• Pytorch 2.6.0+cu124
• Datasets 4.0.0
• Tokenizers 0.21.1