vit-base-patch16-224-RU9-24

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

• Loss: 0.5081
• Accuracy: 0.8431

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: 5.5e-05
• train_batch_size: 32
• eval_batch_size: 32
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 128
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.05
• num_epochs: 24

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	1.0	8	1.3401	0.5098
1.3685	2.0	16	1.2193	0.5686
1.2413	3.0	24	1.1150	0.5882
1.1126	4.0	32	0.9957	0.7059
0.9285	5.0	40	0.8976	0.6863
0.9285	6.0	48	0.8580	0.6863
0.7793	7.0	56	0.8426	0.7647
0.6291	8.0	64	0.7899	0.6863
0.5401	9.0	72	0.7169	0.7255
0.4358	10.0	80	0.7505	0.7255
0.4358	11.0	88	0.8077	0.7059
0.3901	12.0	96	0.6803	0.7647
0.3033	13.0	104	0.6483	0.7647
0.267	14.0	112	0.6451	0.7451
0.2212	15.0	120	0.6119	0.7647
0.2212	16.0	128	0.6150	0.8039
0.2206	17.0	136	0.6270	0.7843
0.2285	18.0	144	0.6181	0.7647
0.1741	19.0	152	0.5081	0.8431
0.1708	20.0	160	0.5502	0.8235
0.1708	21.0	168	0.5689	0.8039
0.16	22.0	176	0.5137	0.8235
0.1567	23.0	184	0.5207	0.8431
0.1616	24.0	192	0.5375	0.8235

Framework versions

• Transformers 4.36.2
• Pytorch 2.1.2+cu118
• Datasets 2.16.1
• Tokenizers 0.15.0