vit-base-patch16-224-in21k-rotated-dungeons-v002

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

• Loss: 1.0084
• Accuracy: 0.8571

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: 1e-05
• train_batch_size: 4
• eval_batch_size: 8
• seed: 1024
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 12
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	1.0	12	1.1281	0.8571
No log	2.0	24	1.1248	0.8571
No log	3.0	36	1.0930	0.8571
No log	4.0	48	1.1040	0.8214
No log	5.0	60	1.0646	0.8214
No log	6.0	72	1.0540	0.8214
No log	7.0	84	1.0309	0.8571
No log	8.0	96	1.0274	0.8571
No log	9.0	108	1.0155	0.8571
No log	10.0	120	1.0079	0.8571
No log	11.0	132	1.0175	0.8571
No log	12.0	144	1.0029	0.8571

Framework versions

• Transformers 4.44.2
• Pytorch 2.5.0+cu121
• Datasets 3.1.0
• Tokenizers 0.19.1