vit-lr-0.001

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

• Loss: 0.7065
• Accuracy: 0.7334
• Precision: 0.7056
• Recall: 0.7334
• F1: 0.7133

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: 0.001
• train_batch_size: 16
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine
• num_epochs: 100
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision	Recall	F1
1.0781	1.0	321	0.9802	0.6709	0.4973	0.6709	0.5438
0.9514	2.0	642	0.8989	0.6689	0.5299	0.6689	0.5529
0.9232	3.0	963	0.9391	0.6702	0.5516	0.6702	0.6020
0.9191	4.0	1284	0.9349	0.6803	0.5261	0.6803	0.5764
0.8919	5.0	1605	0.8966	0.6827	0.5608	0.6827	0.6014
0.8521	6.0	1926	0.8330	0.6928	0.6099	0.6928	0.6337
0.8304	7.0	2247	0.8980	0.6786	0.5849	0.6786	0.5701
0.7971	8.0	2568	0.8211	0.6834	0.6957	0.6834	0.6662
0.7881	9.0	2889	0.8262	0.6963	0.6476	0.6963	0.6051
0.7796	10.0	3210	0.8119	0.7049	0.6523	0.7049	0.6407
0.7535	11.0	3531	0.7961	0.6976	0.7043	0.6976	0.6933
0.7404	12.0	3852	0.7899	0.7063	0.7079	0.7063	0.7019
0.7455	13.0	4173	0.8245	0.6963	0.5619	0.6963	0.6146
0.747	14.0	4494	0.7627	0.7074	0.7085	0.7074	0.6971
0.7263	15.0	4815	0.7411	0.7243	0.7074	0.7243	0.6857
0.7203	16.0	5136	0.7484	0.7098	0.6399	0.7098	0.6503
0.7194	17.0	5457	0.7248	0.7191	0.6724	0.7191	0.6771
0.7029	18.0	5778	0.7470	0.7174	0.6885	0.7174	0.6898
0.7129	19.0	6099	0.7173	0.7243	0.6964	0.7243	0.7037
0.6895	20.0	6420	0.7423	0.7205	0.6830	0.7205	0.6893
0.7066	21.0	6741	0.7528	0.7146	0.6887	0.7146	0.6811
0.6895	22.0	7062	0.7132	0.7302	0.7057	0.7302	0.7102
0.6918	23.0	7383	0.7690	0.7115	0.7080	0.7115	0.6986
0.6818	24.0	7704	0.7065	0.7334	0.7056	0.7334	0.7133
0.6538	25.0	8025	0.7207	0.7230	0.7155	0.7230	0.7185
0.655	26.0	8346	0.7191	0.7268	0.7035	0.7268	0.7090
0.6612	27.0	8667	0.7314	0.7195	0.7167	0.7195	0.6979
0.6406	28.0	8988	0.7939	0.7247	0.6993	0.7247	0.7037
0.64	29.0	9309	0.7089	0.7365	0.7179	0.7365	0.7168
0.637	30.0	9630	0.7512	0.7313	0.6939	0.7313	0.6993
0.622	31.0	9951	0.7483	0.7288	0.6973	0.7288	0.7064
0.6193	32.0	10272	0.7290	0.7198	0.7197	0.7198	0.7175
0.6058	33.0	10593	0.7299	0.7309	0.7053	0.7309	0.7095
0.6047	34.0	10914	0.7086	0.7292	0.7296	0.7292	0.7233

Framework versions

• Transformers 4.39.0.dev0
• Pytorch 2.2.1+cu121
• Datasets 2.18.0
• Tokenizers 0.15.2