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

This model was trained from scratch on the imagefolder dataset. It achieves the following results on the evaluation set:

• Loss: 0.0603
• Accuracy: 0.9816

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: 128
• eval_batch_size: 128
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 512
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 100

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
1.0105	0.99	78	0.9788	0.5251
0.8749	2.0	157	0.8167	0.6323
0.7352	2.99	235	0.6498	0.7179
0.647	4.0	314	0.5619	0.7630
0.5949	4.99	392	0.5068	0.7892
0.5773	6.0	471	0.4673	0.8058
0.5401	6.99	549	0.4277	0.8211
0.5116	8.0	628	0.4089	0.8314
0.4751	8.99	706	0.3891	0.8399
0.4659	10.0	785	0.3599	0.8574
0.4457	10.99	863	0.3493	0.8592
0.4021	12.0	942	0.3180	0.8751
0.3925	12.99	1020	0.2824	0.8863
0.3737	14.0	1099	0.2686	0.8960
0.3573	14.99	1177	0.2509	0.8978
0.3463	16.0	1256	0.2214	0.9101
0.3174	16.99	1334	0.2233	0.9143
0.3029	18.0	1413	0.2065	0.9222
0.27	18.99	1491	0.1900	0.9287
0.2796	20.0	1570	0.1843	0.9294
0.2649	20.99	1648	0.1811	0.9291
0.2561	22.0	1727	0.1679	0.9345
0.2384	22.99	1805	0.1579	0.9401
0.2415	24.0	1884	0.1414	0.9462
0.2282	24.99	1962	0.1478	0.9439
0.2345	26.0	2041	0.1523	0.9419
0.2196	26.99	2119	0.1240	0.9529
0.2118	28.0	2198	0.1228	0.9502
0.1977	28.99	2276	0.1192	0.9565
0.1994	30.0	2355	0.1208	0.9529
0.1959	30.99	2433	0.1189	0.9563
0.1946	32.0	2512	0.1106	0.9617
0.1777	32.99	2590	0.1048	0.9587
0.1895	34.0	2669	0.0965	0.9643
0.1779	34.99	2747	0.1081	0.9623
0.1717	36.0	2826	0.1155	0.9590
0.1715	36.99	2904	0.1006	0.9630
0.1748	38.0	2983	0.0894	0.9695
0.1677	38.99	3061	0.1037	0.9639
0.1723	40.0	3140	0.1009	0.9637
0.1552	40.99	3218	0.0867	0.9697
0.1658	42.0	3297	0.0860	0.9702
0.1531	42.99	3375	0.0921	0.9673
0.1593	44.0	3454	0.0941	0.9646
0.1568	44.99	3532	0.0879	0.9686
0.1621	46.0	3611	0.0929	0.9693
0.1602	46.99	3689	0.0777	0.9744
0.1397	48.0	3768	0.0821	0.9733
0.1475	48.99	3846	0.0810	0.9711
0.1452	50.0	3925	0.0943	0.9697
0.148	50.99	4003	0.0943	0.9704
0.1392	52.0	4082	0.0869	0.9695
0.1389	52.99	4160	0.0876	0.9738
0.1295	54.0	4239	0.0716	0.9749
0.1394	54.99	4317	0.0760	0.9751
0.1354	56.0	4396	0.0771	0.9726
0.1303	56.99	4474	0.0738	0.9738
0.1274	58.0	4553	0.0834	0.9724
0.1276	58.99	4631	0.0801	0.9747
0.1372	60.0	4710	0.0667	0.9762
0.1417	60.99	4788	0.0747	0.9738
0.1249	62.0	4867	0.0735	0.9744
0.1212	62.99	4945	0.0665	0.9780
0.1218	64.0	5024	0.0680	0.9771
0.1193	64.99	5102	0.0679	0.9760
0.1148	66.0	5181	0.0685	0.9774
0.1242	66.99	5259	0.0647	0.9800
0.1167	68.0	5338	0.0646	0.9783
0.117	68.99	5416	0.0763	0.9765
0.1153	70.0	5495	0.0720	0.9753
0.12	70.99	5573	0.0717	0.9771
0.1054	72.0	5652	0.0677	0.9767
0.1183	72.99	5730	0.0741	0.9756
0.1082	74.0	5809	0.0676	0.9787
0.1088	74.99	5887	0.0700	0.9751
0.1125	76.0	5966	0.0663	0.9785
0.1099	76.99	6044	0.0622	0.9789
0.1128	78.0	6123	0.0660	0.9794
0.1178	78.99	6201	0.0699	0.9780
0.1129	80.0	6280	0.0605	0.9794
0.1013	80.99	6358	0.0694	0.9778
0.1078	82.0	6437	0.0652	0.9783
0.0994	82.99	6515	0.0604	0.9812
0.1093	84.0	6594	0.0600	0.9805
0.1039	84.99	6672	0.0646	0.9787
0.0963	86.0	6751	0.0654	0.9789
0.1053	86.99	6829	0.0627	0.9803
0.0982	88.0	6908	0.0619	0.9800
0.0944	88.99	6986	0.0607	0.9796
0.0959	90.0	7065	0.0661	0.9800
0.101	90.99	7143	0.0642	0.9809
0.1095	92.0	7222	0.0607	0.9807
0.1079	92.99	7300	0.0610	0.9803
0.1153	94.0	7379	0.0632	0.9798
0.1022	94.99	7457	0.0618	0.9812
0.1079	96.0	7536	0.0606	0.9809
0.0942	96.99	7614	0.0612	0.9800
0.0927	98.0	7693	0.0598	0.9809
0.1032	98.99	7771	0.0604	0.9814
0.0925	99.36	7800	0.0603	0.9816

Framework versions

• Transformers 4.38.1
• Pytorch 2.1.2+cu121
• Datasets 2.17.1
• Tokenizers 0.15.2