vit-beta1-0.88

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: 0.4922
• Accuracy: 0.8523
• Precision: 0.8564
• Recall: 0.8523
• F1: 0.8539

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=(0.88,0.999) 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
1.713	1.0	321	0.9541	0.6987	0.6717	0.6987	0.6282
1.1578	2.0	642	0.7858	0.6956	0.7680	0.6956	0.7157
1.035	3.0	963	0.7133	0.7282	0.7631	0.7282	0.6990
0.9749	4.0	1284	0.5555	0.7712	0.7960	0.7712	0.7759
0.9392	5.0	1605	0.6676	0.7642	0.7943	0.7642	0.7721
0.992	6.0	1926	0.6117	0.7725	0.7958	0.7725	0.7767
0.8857	7.0	2247	0.6220	0.7562	0.7969	0.7562	0.7664
0.7885	8.0	2568	0.7200	0.7042	0.8134	0.7042	0.7319
0.7404	9.0	2889	0.5796	0.7798	0.7973	0.7798	0.7801
0.6687	10.0	3210	0.5593	0.7757	0.8340	0.7757	0.7910
0.6173	11.0	3531	0.5573	0.8017	0.8328	0.8017	0.8090
0.5531	12.0	3852	0.5931	0.7902	0.8270	0.7902	0.8024
0.4717	13.0	4173	0.6534	0.8221	0.8265	0.8221	0.8126
0.4553	14.0	4494	0.5501	0.8329	0.8489	0.8329	0.8367
0.3776	15.0	4815	0.6037	0.8193	0.8392	0.8193	0.8242
0.3435	16.0	5136	0.6238	0.8093	0.8367	0.8093	0.8176
0.2983	17.0	5457	0.4922	0.8523	0.8564	0.8523	0.8539
0.259	18.0	5778	0.5820	0.8356	0.8512	0.8356	0.8406
0.2105	19.0	6099	0.6471	0.8270	0.8536	0.8270	0.8349
0.2268	20.0	6420	0.6060	0.8519	0.8551	0.8519	0.8527
0.155	21.0	6741	0.5968	0.8693	0.8649	0.8693	0.8663
0.1226	22.0	7062	0.6047	0.8682	0.8661	0.8682	0.8668
0.1267	23.0	7383	0.6395	0.8682	0.8617	0.8682	0.8629
0.1184	24.0	7704	0.6441	0.8693	0.8660	0.8693	0.8660
0.1034	25.0	8025	0.6301	0.8689	0.8671	0.8689	0.8672
0.0907	26.0	8346	0.6288	0.8717	0.8674	0.8717	0.8687
0.1358	27.0	8667	0.6250	0.8727	0.8682	0.8727	0.8698

Framework versions

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