model

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

• Loss: 1.4897
• Accuracy: 0.6

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: 5e-05
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 25

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	1.0	80	1.7001	0.325
No log	2.0	160	1.4642	0.4875
No log	3.0	240	1.3522	0.4625
No log	4.0	320	1.3493	0.4688
No log	5.0	400	1.2052	0.55
No log	6.0	480	1.2267	0.5563
1.2917	7.0	560	1.1744	0.6062
1.2917	8.0	640	1.2969	0.5437
1.2917	9.0	720	1.2519	0.5687
1.2917	10.0	800	1.3108	0.5125
1.2917	11.0	880	1.2725	0.5875
1.2917	12.0	960	1.3437	0.55
0.5002	13.0	1040	1.3790	0.5375
0.5002	14.0	1120	1.3432	0.625
0.5002	15.0	1200	1.4395	0.55
0.5002	16.0	1280	1.3672	0.5875
0.5002	17.0	1360	1.3928	0.575
0.5002	18.0	1440	1.3016	0.5875
0.2523	19.0	1520	1.4815	0.5625
0.2523	20.0	1600	1.3394	0.6062
0.2523	21.0	1680	1.3450	0.5938
0.2523	22.0	1760	1.3924	0.6312
0.2523	23.0	1840	1.4664	0.5813
0.2523	24.0	1920	1.2635	0.65
0.1723	25.0	2000	1.4154	0.5625

Framework versions

• Transformers 4.35.2
• Pytorch 2.1.0+cu121
• Datasets 2.17.0
• Tokenizers 0.15.1