vit-base-patch16-224-isic248

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

• Loss: 1.4425
• Accuracy: 0.6

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: 2
• eval_batch_size: 2
• seed: 1
• gradient_accumulation_steps: 4
• total_train_batch_size: 8
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 30

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
2.4757	0.9888	22	2.2018	0.15
2.0035	1.9775	44	1.9165	0.35
1.5959	2.9663	66	1.6904	0.35
1.1695	4.0	89	1.5546	0.45
0.8869	4.9888	111	1.5021	0.45
0.6398	5.9775	133	1.5211	0.5
0.5405	6.9663	155	1.4225	0.55
0.3626	8.0	178	1.3762	0.45
0.2971	8.9888	200	1.5558	0.55
0.2324	9.9775	222	1.4037	0.55
0.169	10.9663	244	1.4982	0.55
0.1625	12.0	267	1.3437	0.55
0.1088	12.9888	289	1.2965	0.55
0.1924	13.9775	311	1.2558	0.5
0.124	14.9663	333	1.4687	0.5
0.0901	16.0	356	1.5054	0.6
0.0641	16.9888	378	1.4898	0.5
0.1093	17.9775	400	1.3880	0.55
0.1307	18.9663	422	1.4609	0.45
0.1268	20.0	445	1.5605	0.6
0.0416	20.9888	467	1.5644	0.55
0.0538	21.9775	489	1.5292	0.6
0.109	22.9663	511	1.4049	0.6
0.0633	24.0	534	1.5017	0.55
0.031	24.9888	556	1.4835	0.55
0.0926	25.9775	578	1.4584	0.6
0.067	26.9663	600	1.5057	0.65
0.0344	28.0	623	1.4680	0.6
0.0806	28.9888	645	1.4399	0.6
0.0329	29.6629	660	1.4425	0.6

Framework versions

• Transformers 4.41.2
• Pytorch 2.1.0
• Datasets 2.19.2
• Tokenizers 0.19.1