test_trainer

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: 0.8643
• Accuracy: 0.915

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: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	1.0	125	3.6903	0.517
No log	2.0	250	2.7990	0.553
No log	3.0	375	2.3198	0.57
3.1391	4.0	500	2.0210	0.632
3.1391	5.0	625	1.8298	0.638
3.1391	6.0	750	1.6753	0.683
3.1391	7.0	875	1.5446	0.708
1.7309	8.0	1000	1.4338	0.751
1.7309	9.0	1125	1.3318	0.777
1.7309	10.0	1250	1.2387	0.807
1.7309	11.0	1375	1.1828	0.806
1.2855	12.0	1500	1.1052	0.843
1.2855	13.0	1625	1.0620	0.862
1.2855	14.0	1750	1.0029	0.87
1.2855	15.0	1875	0.9611	0.895
1.0212	16.0	2000	0.9314	0.905
1.0212	17.0	2125	0.9041	0.905
1.0212	18.0	2250	0.8840	0.913
1.0212	19.0	2375	0.8730	0.921
0.8953	20.0	2500	0.8639	0.92

Framework versions

• Transformers 4.37.1
• Pytorch 2.1.0+cu121
• Datasets 2.16.1
• Tokenizers 0.15.1