vit-beta2-0.99

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: 0.5248
• Accuracy: 0.8623
• Precision: 0.8584
• Recall: 0.8623
• F1: 0.8596

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: 0.0001
• train_batch_size: 16
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.99) and epsilon=1e-08
• lr_scheduler_type: cosine
• lr_scheduler_warmup_steps: 1733
• num_epochs: 100
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision	Recall	F1
1.7217	1.0	321	1.0190	0.7035	0.6635	0.7035	0.6527
1.1622	2.0	642	0.7386	0.7056	0.7576	0.7056	0.7192
1.0368	3.0	963	0.6550	0.7517	0.7564	0.7517	0.7214
0.9653	4.0	1284	0.5641	0.7843	0.7948	0.7843	0.7863
0.9272	5.0	1605	0.7957	0.6768	0.7959	0.6768	0.7035
0.9878	6.0	1926	0.5809	0.7871	0.8062	0.7871	0.7904
0.872	7.0	2247	0.6815	0.7216	0.8081	0.7216	0.7442
0.7998	8.0	2568	0.6104	0.7559	0.8143	0.7559	0.7723
0.733	9.0	2889	0.5296	0.8148	0.8254	0.8148	0.8172
0.6957	10.0	3210	0.5797	0.7996	0.8322	0.7996	0.8052
0.6271	11.0	3531	0.5926	0.7933	0.8343	0.7933	0.8058
0.5614	12.0	3852	0.5879	0.7920	0.8384	0.7920	0.8060
0.4576	13.0	4173	0.6665	0.8138	0.8312	0.8138	0.8028
0.4645	14.0	4494	0.5515	0.8294	0.8470	0.8294	0.8329
0.3913	15.0	4815	0.5474	0.8225	0.8466	0.8225	0.8288
0.3693	16.0	5136	0.5769	0.8235	0.8464	0.8235	0.8308
0.2794	17.0	5457	0.5328	0.8509	0.8571	0.8509	0.8516
0.2677	18.0	5778	0.5248	0.8623	0.8584	0.8623	0.8596
0.2104	19.0	6099	0.6284	0.8433	0.8572	0.8433	0.8473
0.2459	20.0	6420	0.6137	0.8544	0.8596	0.8544	0.8555
0.1769	21.0	6741	0.5960	0.8637	0.8573	0.8637	0.8566
0.1294	22.0	7062	0.5844	0.8700	0.8687	0.8700	0.8687
0.1597	23.0	7383	0.6580	0.8665	0.8604	0.8665	0.8589
0.1227	24.0	7704	0.6226	0.8731	0.8720	0.8731	0.8712
0.1054	25.0	8025	0.6198	0.8752	0.8728	0.8752	0.8721
0.0945	26.0	8346	0.6050	0.8793	0.8757	0.8793	0.8764
0.1242	27.0	8667	0.6078	0.8828	0.8788	0.8828	0.8798
0.0819	28.0	8988	0.6190	0.8797	0.8748	0.8797	0.8756

Framework versions

• Transformers 4.40.0.dev0
• Pytorch 2.2.1+cu121
• Datasets 2.18.0
• Tokenizers 0.15.2