vit-large-patch16-224-in21k-dungeon-geo-morphs-denoised-04Dec24-001

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: 0.0863
• Accuracy: 0.9697

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
1.3958	4.0	10	0.9619	0.8384
0.4596	8.0	20	0.3876	0.9333
0.1012	12.0	30	0.1928	0.9616
0.022	16.0	40	0.1181	0.9636
0.0066	20.0	50	0.0936	0.9677
0.0036	24.0	60	0.0863	0.9697
0.0028	28.0	70	0.0848	0.9697
0.0025	32.0	80	0.0826	0.9697

Framework versions

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