Boya1_SGD_1-e3_20Epoch_Deit-tiny-patch16_fold4

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: 1.7238
• Accuracy: 0.4238

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
2.4152	1.0	923	2.4593	0.2073
2.4387	2.0	1846	2.2993	0.2512
2.1969	3.0	2769	2.1607	0.3133
2.0455	4.0	3692	2.0589	0.3320
1.8171	5.0	4615	1.9845	0.3585
1.8796	6.0	5538	1.9302	0.3656
1.8281	7.0	6461	1.8840	0.3816
1.7455	8.0	7384	1.8500	0.3883
1.7072	9.0	8307	1.8232	0.4003
1.7401	10.0	9230	1.8005	0.4046
1.8157	11.0	10153	1.7845	0.4114
1.796	12.0	11076	1.7690	0.4114
1.7335	13.0	11999	1.7588	0.4122
1.6292	14.0	12922	1.7473	0.4190
1.7133	15.0	13845	1.7397	0.4222
1.7521	16.0	14768	1.7345	0.4195
1.8322	17.0	15691	1.7291	0.4244
1.7763	18.0	16614	1.7260	0.4244
1.5996	19.0	17537	1.7248	0.4225
1.6259	20.0	18460	1.7238	0.4238

Framework versions

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