vit_4090_1_9

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

• Loss: 0.1320
• Accuracy: 0.9724
• Precision: 0.9730
• Recall: 0.9724
• F1: 0.9713

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: 0.0001
• train_batch_size: 24
• eval_batch_size: 4
• seed: 42
• 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: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision	Recall	F1
0.5464	1.0	177	0.2534	0.9257	0.9311	0.9257	0.9243
0.2416	2.0	354	0.1512	0.9575	0.9579	0.9575	0.9555
0.1362	3.0	531	0.1294	0.9745	0.9749	0.9745	0.9740
0.0913	4.0	708	0.1155	0.9660	0.9664	0.9660	0.9636
0.0649	5.0	885	0.1336	0.9724	0.9729	0.9724	0.9714
0.0482	6.0	1062	0.1733	0.9660	0.9674	0.9660	0.9652
0.0307	7.0	1239	0.1847	0.9745	0.9753	0.9745	0.9736
0.0356	8.0	1416	0.1455	0.9703	0.9708	0.9703	0.9690
0.015	9.0	1593	0.1237	0.9745	0.9779	0.9745	0.9743
0.0079	10.0	1770	0.1545	0.9724	0.9739	0.9724	0.9722
0.0086	11.0	1947	0.1669	0.9745	0.9755	0.9745	0.9735
0.0059	12.0	2124	0.1449	0.9682	0.9697	0.9682	0.9683
0.0067	13.0	2301	0.1734	0.9724	0.9735	0.9724	0.9718
0.0009	14.0	2478	0.1419	0.9745	0.9749	0.9745	0.9739
0.0004	15.0	2655	0.1510	0.9745	0.9750	0.9745	0.9737
0.0007	16.0	2832	0.1676	0.9724	0.9730	0.9724	0.9713
0.0011	17.0	3009	0.1225	0.9724	0.9730	0.9724	0.9713
0.0015	18.0	3186	0.1345	0.9682	0.9691	0.9682	0.9675
0.0001	19.0	3363	0.1263	0.9724	0.9730	0.9724	0.9713
0.0001	20.0	3540	0.1320	0.9724	0.9730	0.9724	0.9713

Framework versions

• Transformers 4.49.0
• Pytorch 2.5.1
• Datasets 3.2.0
• Tokenizers 0.21.1