swin-tiny-patch4-window7-224-finetuned-main-gpu-20e-final

This model is a fine-tuned version of microsoft/swin-tiny-patch4-window7-224 on the imagefolder dataset. It achieves the following results on the evaluation set:

• Loss: 0.0251
• Accuracy: 0.9917

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: 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: 0.1
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
0.5767	1.0	551	0.5565	0.7463
0.3985	2.0	1102	0.3165	0.8711
0.2988	3.0	1653	0.1835	0.9293
0.2449	4.0	2204	0.1150	0.9572
0.2037	5.0	2755	0.0993	0.9632
0.1646	6.0	3306	0.0750	0.9717
0.1995	7.0	3857	0.0610	0.9776
0.1659	8.0	4408	0.0485	0.9815
0.1449	9.0	4959	0.0505	0.9821
0.1315	10.0	5510	0.0444	0.9843
0.102	11.0	6061	0.0440	0.9838
0.1039	12.0	6612	0.0359	0.9870
0.0798	13.0	7163	0.0393	0.9869
0.1033	14.0	7714	0.0343	0.9890
0.078	15.0	8265	0.0298	0.9902
0.0765	16.0	8816	0.0299	0.9901
0.0769	17.0	9367	0.0275	0.9908
0.0751	18.0	9918	0.0271	0.9910
0.0822	19.0	10469	0.0251	0.9917
0.0756	20.0	11020	0.0254	0.9913

Framework versions

• Transformers 4.26.1
• Pytorch 1.13.1+cu116
• Datasets 2.10.1
• Tokenizers 0.13.2