swin-tiny-patch4-window7-224-finetuned-eurosat_DATA6_20240305

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

• Loss: 0.0731
• Accuracy: 0.9747

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
1.0432	0.99	78	1.0627	0.4531
0.9424	2.0	157	1.0076	0.5103
0.8681	2.99	235	0.9070	0.5778
0.7533	4.0	314	0.7472	0.6731
0.655	4.99	392	0.6276	0.7339
0.6233	6.0	471	0.5866	0.7572
0.592	6.99	549	0.5434	0.7711
0.5502	8.0	628	0.4688	0.8045
0.5192	8.99	706	0.4532	0.8139
0.4923	10.0	785	0.4253	0.8303
0.487	10.99	863	0.4004	0.8365
0.432	12.0	942	0.3639	0.8531
0.4194	12.99	1020	0.3321	0.8675
0.4031	14.0	1099	0.3658	0.8545
0.3803	14.99	1177	0.2912	0.8861
0.3742	16.0	1256	0.2719	0.8906
0.3767	16.99	1334	0.2803	0.8874
0.3281	18.0	1413	0.2732	0.8975
0.313	18.99	1491	0.2466	0.9065
0.3147	20.0	1570	0.2308	0.9096
0.3053	20.99	1648	0.2295	0.9170
0.2784	22.0	1727	0.2241	0.9170
0.2788	22.99	1805	0.2084	0.9240
0.2665	24.0	1884	0.2074	0.9197
0.2528	24.99	1962	0.1779	0.9339
0.2518	26.0	2041	0.1782	0.9359
0.2502	26.99	2119	0.1645	0.9415
0.2387	28.0	2198	0.1856	0.9312
0.2459	28.99	2276	0.1496	0.9430
0.2122	30.0	2355	0.1503	0.9469
0.2139	30.99	2433	0.1598	0.9424
0.2065	32.0	2512	0.1484	0.9448
0.2052	32.99	2590	0.1613	0.9417
0.2104	34.0	2669	0.1353	0.9561
0.2068	34.99	2747	0.1461	0.9464
0.2009	36.0	2826	0.1307	0.9529
0.1862	36.99	2904	0.1375	0.9509
0.1925	38.0	2983	0.1216	0.9556
0.192	38.99	3061	0.1215	0.9549
0.1974	40.0	3140	0.1301	0.9587
0.1858	40.99	3218	0.1156	0.9565
0.1802	42.0	3297	0.1105	0.9635
0.1828	42.99	3375	0.1394	0.9565
0.1767	44.0	3454	0.1142	0.9612
0.1852	44.99	3532	0.1097	0.9630
0.1717	46.0	3611	0.1147	0.9596
0.1799	46.99	3689	0.1222	0.9610
0.166	48.0	3768	0.1133	0.9608
0.1638	48.99	3846	0.1058	0.9646
0.1524	50.0	3925	0.1045	0.9626
0.1581	50.99	4003	0.1065	0.9632
0.1616	52.0	4082	0.0996	0.9691
0.158	52.99	4160	0.1048	0.9648
0.1466	54.0	4239	0.1047	0.9639
0.1617	54.99	4317	0.0982	0.9650
0.1624	56.0	4396	0.1010	0.9643
0.1465	56.99	4474	0.0999	0.9657
0.1368	58.0	4553	0.0977	0.9648
0.1445	58.99	4631	0.1106	0.9630
0.1368	60.0	4710	0.0985	0.9661
0.1401	60.99	4788	0.0913	0.9686
0.1447	62.0	4867	0.0888	0.9688
0.1356	62.99	4945	0.0894	0.9704
0.1358	64.0	5024	0.0851	0.9700
0.1345	64.99	5102	0.0920	0.9709
0.1297	66.0	5181	0.0858	0.9704
0.134	66.99	5259	0.0875	0.9709
0.1252	68.0	5338	0.0954	0.9688
0.139	68.99	5416	0.0869	0.9733
0.1341	70.0	5495	0.0816	0.9691
0.1263	70.99	5573	0.0835	0.9726
0.1247	72.0	5652	0.0858	0.9686
0.122	72.99	5730	0.0891	0.9688
0.118	74.0	5809	0.0829	0.9715
0.1236	74.99	5887	0.0743	0.9751
0.1305	76.0	5966	0.0882	0.9715
0.1219	76.99	6044	0.0826	0.9722
0.1184	78.0	6123	0.0838	0.9722
0.124	78.99	6201	0.0819	0.9731
0.118	80.0	6280	0.0859	0.9709
0.1082	80.99	6358	0.0816	0.9726
0.1195	82.0	6437	0.0783	0.9722
0.1101	82.99	6515	0.0792	0.9747
0.1233	84.0	6594	0.0759	0.9753
0.1181	84.99	6672	0.0783	0.9722
0.1153	86.0	6751	0.0817	0.9715
0.1116	86.99	6829	0.0856	0.9733
0.1066	88.0	6908	0.0788	0.9758
0.1047	88.99	6986	0.0771	0.9735
0.1109	90.0	7065	0.0733	0.9744
0.1125	90.99	7143	0.0748	0.9747
0.1166	92.0	7222	0.0757	0.9738
0.1021	92.99	7300	0.0778	0.9740
0.1089	94.0	7379	0.0741	0.9753
0.1061	94.99	7457	0.0718	0.9765
0.1144	96.0	7536	0.0728	0.9747
0.1146	96.99	7614	0.0737	0.9747
0.1097	98.0	7693	0.0727	0.9749
0.1139	98.99	7771	0.0731	0.9747
0.1025	99.36	7800	0.0731	0.9747

Framework versions

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