Boya1_SGD_1-e3_20Epoch_Deit-tiny-patch16_fold3

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

• Loss: 1.7155
• Accuracy: 0.4288

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: 0.001
• train_batch_size: 16
• eval_batch_size: 16
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
2.4106	1.0	923	2.4578	0.2002
2.3587	2.0	1846	2.2972	0.2516
2.1274	3.0	2769	2.1627	0.3055
2.1583	4.0	3692	2.0604	0.3279
1.9036	5.0	4615	1.9842	0.3458
1.7721	6.0	5538	1.9243	0.3582
1.9867	7.0	6461	1.8782	0.3726
1.8532	8.0	7384	1.8428	0.3891
1.8503	9.0	8307	1.8165	0.4004
1.79	10.0	9230	1.7943	0.4037
1.7717	11.0	10153	1.7761	0.4091
1.7696	12.0	11076	1.7613	0.4148
1.7298	13.0	11999	1.7507	0.4191
1.7468	14.0	12922	1.7401	0.4210
1.6085	15.0	13845	1.7322	0.4229
1.7188	16.0	14768	1.7257	0.4278
1.7307	17.0	15691	1.7212	0.4259
1.5257	18.0	16614	1.7177	0.4275
1.6729	19.0	17537	1.7160	0.4294
1.7293	20.0	18460	1.7155	0.4288

Framework versions

• Transformers 4.40.1
• Pytorch 2.1.0
• Datasets 2.19.0
• Tokenizers 0.19.1