vit-wonders-model

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

• Loss: 0.0541
• Accuracy: 0.9959
• F1: 0.9957

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: 0.0002
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• 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: linear
• num_epochs: 3
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1
0.2904	0.4115	100	0.3038	0.9588	0.9630
0.2357	0.8230	200	0.1637	0.9630	0.9660
0.052	1.2346	300	0.0541	0.9959	0.9957
0.0229	1.6461	400	0.0763	0.9877	0.9836
0.025	2.0576	500	0.0630	0.9918	0.9911

Framework versions

• Transformers 4.55.2
• Pytorch 2.6.0
• Datasets 3.3.2
• Tokenizers 0.21.0