vit-dropout-v11

This model is a fine-tuned version of google/vit-base-patch16-224-in21k on the SkinCancerClassification dataset. It achieves the following results on the evaluation set:

• Loss: 0.5383
• Accuracy: 0.8508

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

• dropout: 0.35

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: 500
• num_epochs: 16
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
0.5972	1.56	500	0.9045	0.7160
0.5053	3.12	1000	0.6760	0.7566
0.5003	4.67	1500	0.5776	0.7909
0.4323	6.23	2000	0.5317	0.8084
0.417	7.79	2500	0.5187	0.8277
0.2751	9.35	3000	0.5660	0.8346
0.3664	10.9	3500	0.5247	0.8371
0.2595	12.46	4000	0.5358	0.8546
0.2009	14.02	4500	0.5289	0.8508
0.1366	15.58	5000	0.5383	0.8508

Framework versions

• Transformers 4.38.1
• Pytorch 2.1.0+cu121
• Datasets 2.17.1
• Tokenizers 0.15.2