vit-base-patch16-224-in21k-finetuned-earlystop

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

• Loss: 0.1972
• Accuracy: 0.9375

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: 5e-05
• train_batch_size: 32
• eval_batch_size: 32
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 128
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: cosine
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
0.5989	0.9816	40	0.6929	0.5
0.3542	1.9877	81	0.5951	0.6875
0.2495	2.9939	122	0.5182	0.75
0.1553	4.0	163	0.7023	0.625
0.1806	4.9816	203	0.3825	0.8125
0.1509	5.9877	244	0.1972	0.9375
0.1771	6.9939	285	0.6752	0.625
0.1372	8.0	326	0.4901	0.6875
0.1698	8.9816	366	0.2187	0.875

Framework versions

• Transformers 4.46.3
• Pytorch 2.5.1+cu121
• Datasets 3.1.0
• Tokenizers 0.20.3