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

This model is a fine-tuned version of microsoft/swin-tiny-patch4-window7-224 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.0846
• F1: 0.5965
• Roc Auc: 0.7500
• Accuracy: 0.3659

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: 40

Training results

Training Loss	Epoch	Step	Validation Loss	F1	Roc Auc	Accuracy
No log	0.91	8	0.6685	0.0992	0.5341	0.0
0.7041	1.94	17	0.4597	0.1245	0.5440	0.0
0.5413	2.97	26	0.1970	0.0287	0.5067	0.0081
0.2531	4.0	35	0.1487	0.0099	0.5022	0.0
0.15	4.91	43	0.1465	0.0566	0.5142	0.0407
0.145	5.94	52	0.1433	0.1166	0.5312	0.0569
0.138	6.97	61	0.1412	0.2140	0.5629	0.0976
0.1374	8.0	70	0.1377	0.2698	0.5827	0.1138
0.1374	8.91	78	0.1319	0.2410	0.5726	0.1057
0.1309	9.94	87	0.1284	0.3100	0.6014	0.1382
0.1256	10.97	96	0.1228	0.2667	0.5824	0.1220
0.1196	12.0	105	0.1201	0.3500	0.6186	0.1463
0.116	12.91	113	0.1169	0.3732	0.6286	0.1707
0.1102	13.94	122	0.1137	0.3650	0.6220	0.1951
0.1062	14.97	131	0.1082	0.3843	0.6316	0.2195
0.1019	16.0	140	0.1048	0.4630	0.6751	0.2602
0.1019	16.91	148	0.1033	0.4475	0.6614	0.2602
0.0965	17.94	157	0.1046	0.4890	0.6899	0.2846
0.0935	18.97	166	0.1014	0.4651	0.6711	0.2358
0.0928	20.0	175	0.0998	0.4877	0.6918	0.2520
0.0897	20.91	183	0.0959	0.5145	0.6961	0.2683
0.0843	21.94	192	0.0933	0.5296	0.7080	0.2927
0.0829	22.97	201	0.0919	0.5610	0.7255	0.3171
0.0804	24.0	210	0.0917	0.5644	0.7257	0.3496
0.0804	24.91	218	0.0898	0.6036	0.7505	0.3577
0.0797	25.94	227	0.0886	0.5758	0.7331	0.3333
0.0762	26.97	236	0.0865	0.5740	0.7330	0.3415
0.0757	28.0	245	0.0879	0.5893	0.7429	0.3577
0.0736	28.91	253	0.0866	0.5875	0.7427	0.3415
0.0716	29.94	262	0.0855	0.5910	0.7430	0.3659
0.0722	30.97	271	0.0857	0.5917	0.7452	0.3577
0.0716	32.0	280	0.0864	0.5868	0.7405	0.3415
0.0716	32.91	288	0.0850	0.5917	0.7452	0.3577
0.0701	33.94	297	0.0849	0.5965	0.7500	0.3577
0.0701	34.97	306	0.0844	0.5875	0.7427	0.3496
0.0704	36.0	315	0.0846	0.5982	0.7501	0.3659
0.0695	36.57	320	0.0846	0.5965	0.7500	0.3659

Framework versions

• Transformers 4.34.1
• Pytorch 2.1.0+cu118
• Datasets 2.14.6
• Tokenizers 0.14.1