swin-tiny-patch4-window7-224-finetuned-eurosat

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: 0.2469
• Accuracy: 0.9383

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: 30

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
0.9843	0.99	43	0.8500	0.6948
0.5335	2.0	87	0.5584	0.7825
0.4263	2.99	130	0.4791	0.8117
0.3308	4.0	174	0.4269	0.8344
0.2882	4.99	217	0.3567	0.8636
0.2517	6.0	261	0.3317	0.8701
0.1908	6.99	304	0.3082	0.8815
0.187	8.0	348	0.3230	0.8799
0.1434	8.99	391	0.3323	0.9010
0.1277	10.0	435	0.2489	0.9075
0.156	10.99	478	0.3246	0.8880
0.0781	12.0	522	0.3121	0.9010
0.1001	12.99	565	0.2708	0.9058
0.0892	14.0	609	0.2582	0.9140
0.0644	14.99	652	0.2486	0.9221
0.0689	16.0	696	0.2465	0.9237
0.0547	16.99	739	0.2402	0.9334
0.0597	18.0	783	0.2534	0.9237
0.0512	18.99	826	0.2400	0.9318
0.041	20.0	870	0.2397	0.9286
0.0376	20.99	913	0.2663	0.9269
0.0412	22.0	957	0.3026	0.9221
0.0423	22.99	1000	0.2678	0.9302
0.0266	24.0	1044	0.2510	0.9318
0.0313	24.99	1087	0.2542	0.9334
0.0207	26.0	1131	0.2743	0.9334
0.0292	26.99	1174	0.2614	0.9318
0.0242	28.0	1218	0.2469	0.9383
0.0201	28.99	1261	0.2534	0.9367
0.0354	29.66	1290	0.2525	0.9367

Framework versions

• Transformers 4.37.2
• Pytorch 2.1.0+cu121
• Datasets 2.17.0
• Tokenizers 0.15.1