vit-large-patch16-224-testing-dungeons-lora-23Nov24-009

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

• Loss: 0.1605
• Accuracy: 0.9630

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: 0.005
• 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
• num_epochs: 12
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	0.6667	1	1.4611	0.2963
No log	2.0	3	1.0087	0.5185
No log	2.6667	4	1.5508	0.4074
No log	4.0	6	0.4546	0.8889
No log	4.6667	7	0.4480	0.8148
No log	6.0	9	0.2657	0.8889
0.8959	6.6667	10	0.2433	0.8889
0.8959	8.0	12	0.1605	0.9630

Framework versions

• PEFT 0.13.2
• Transformers 4.46.2
• Pytorch 2.5.1+cu121
• Datasets 3.1.0
• Tokenizers 0.20.3