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: 0.4888
Accuracy: 0.8318
Precision: 0.8450
Recall: 0.8318
F1: 0.8340

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=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: cosine
lr_scheduler_warmup_steps: 1219
num_epochs: 100
mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision	Recall	F1
1.6129	1.0	321	0.8251	0.7223	0.7708	0.7223	0.7354
0.9753	2.0	642	0.7419	0.7115	0.7830	0.7115	0.7320
0.9289	3.0	963	0.5421	0.7968	0.7942	0.7968	0.7901
0.8978	4.0	1284	0.6563	0.7410	0.8112	0.7410	0.7569
0.8271	5.0	1605	0.5586	0.8058	0.8090	0.8058	0.8050
0.7757	6.0	1926	0.5340	0.8065	0.8220	0.8065	0.8113
0.6981	7.0	2247	0.5943	0.7791	0.8207	0.7791	0.7911
0.5954	8.0	2568	0.5836	0.7660	0.8270	0.7660	0.7824
0.5804	9.0	2889	0.4888	0.8318	0.8450	0.8318	0.8340
0.4894	10.0	3210	0.5404	0.8037	0.8459	0.8037	0.8145
0.4924	11.0	3531	0.4997	0.8166	0.8530	0.8166	0.8257
0.38	12.0	3852	0.5669	0.8356	0.8487	0.8356	0.8404
0.3109	13.0	4173	0.5983	0.8467	0.8492	0.8467	0.8415
0.3445	14.0	4494	0.5060	0.8620	0.8643	0.8620	0.8622
0.2457	15.0	4815	0.5823	0.8547	0.8564	0.8547	0.8520
0.2297	16.0	5136	0.5347	0.8620	0.8647	0.8620	0.8628
0.1701	17.0	5457	0.5561	0.8689	0.8665	0.8689	0.8656
0.186	18.0	5778	0.5319	0.8686	0.8688	0.8686	0.8679
0.1258	19.0	6099	0.6001	0.8655	0.8655	0.8655	0.8643

Framework versions

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

sharren/vit-dropout-0.2

vit-dropout-0.2

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

sharren/
vit-dropout-0.2