vit-large-patch16-224-in21k-dungeon-geo-morphs-0-4-30Nov24-003

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

• Loss: 0.1374
• Accuracy: 0.9589

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: 2e-05
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 32
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 40
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
1.3682	3.9091	10	1.0376	0.825
0.5281	7.9091	20	0.5309	0.9196
0.1424	11.9091	30	0.2757	0.9375
0.033	15.9091	40	0.1681	0.9482
0.0093	19.9091	50	0.1374	0.9589
0.0046	23.9091	60	0.1288	0.9589
0.0034	27.9091	70	0.1221	0.9571
0.003	31.9091	80	0.1208	0.9571

Framework versions

• Transformers 4.46.2
• Pytorch 2.5.1+cu121
• Datasets 3.1.0
• Tokenizers 0.20.3