dinov2-base-finetuned-dermnet-lr3-5-0.05wd-csr

This model is a fine-tuned version of facebook/dinov2-base on the imagefolder dataset. It achieves the following results on the evaluation set:

• Loss: 0.8119
• Accuracy: 0.7958

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: 3e-05
• train_batch_size: 32
• eval_batch_size: 32
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 128
• optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: cosine_with_restarts
• lr_scheduler_warmup_ratio: 0.05
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
2.7488	1.0	98	2.5158	0.3722
1.9402	2.0	196	1.7710	0.5170
1.4938	3.0	294	1.4939	0.5996
1.1226	4.0	392	1.3168	0.6256
0.9329	5.0	490	1.1906	0.6705
0.8039	6.0	588	1.0882	0.7067
0.6426	7.0	686	1.1061	0.6930
0.5777	8.0	784	1.0133	0.7227
0.477	9.0	882	0.9681	0.7364
0.3961	10.0	980	0.9402	0.7581
0.3451	11.0	1078	0.9311	0.7509
0.337	12.0	1176	0.8897	0.7661
0.2348	13.0	1274	0.8616	0.7762
0.1992	14.0	1372	0.8241	0.7951
0.182	15.0	1470	0.8312	0.7878
0.1556	16.0	1568	0.8245	0.7857
0.1516	17.0	1666	0.8170	0.7958
0.1569	18.0	1764	0.8202	0.7878
0.1364	19.0	1862	0.8117	0.7951
0.1427	19.8021	1940	0.8119	0.7958

Framework versions

• Transformers 4.50.3
• Pytorch 2.6.0+cu124
• Datasets 3.5.0
• Tokenizers 0.21.1