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: 0.5119
Accuracy: 0.8076
Precision: 0.8207
Recall: 0.8076
F1: 0.8098

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=(0.9,0.995) 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
1.7709	1.0	321	0.9409	0.7039	0.6601	0.7039	0.6317
1.1633	2.0	642	0.7317	0.7372	0.7193	0.7372	0.6970
1.0429	3.0	963	0.6350	0.7625	0.7357	0.7625	0.7240
0.9649	4.0	1284	0.5760	0.7694	0.8038	0.7694	0.7808
0.9051	5.0	1605	0.6441	0.7670	0.7941	0.7670	0.7732
0.9826	6.0	1926	0.5662	0.7850	0.7956	0.7850	0.7892
0.8855	7.0	2247	0.6882	0.7264	0.7937	0.7264	0.7459
0.789	8.0	2568	0.6491	0.7365	0.8089	0.7365	0.7564
0.7192	9.0	2889	0.5119	0.8076	0.8207	0.8076	0.8098
0.7012	10.0	3210	0.5414	0.7975	0.8341	0.7975	0.8077
0.6376	11.0	3531	0.5712	0.7947	0.8332	0.7947	0.8066
0.5412	12.0	3852	0.5661	0.8058	0.8328	0.8058	0.8145
0.4667	13.0	4173	0.6375	0.8190	0.8410	0.8190	0.8179
0.4766	14.0	4494	0.5736	0.8252	0.8508	0.8252	0.8313
0.384	15.0	4815	0.5305	0.8356	0.8415	0.8356	0.8371
0.37	16.0	5136	0.5531	0.8315	0.8499	0.8315	0.8379
0.2809	17.0	5457	0.5174	0.8637	0.8630	0.8637	0.8608
0.2681	18.0	5778	0.5556	0.8478	0.8555	0.8478	0.8504
0.2139	19.0	6099	0.6291	0.8256	0.8518	0.8256	0.8335

Framework versions

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

sharren/vit-beta2-0.995

vit-beta2-0.995

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

sharren/
vit-beta2-0.995