vit-base-patch16-224-finetuned-eurosat-2

This model is a fine-tuned version of google/vit-base-patch16-224 on the imagefolder dataset. It achieves the following results on the evaluation set:

• Loss: 0.0342
• Accuracy: 0.9902
• F1 Macro: 0.9865
• Precision Macro: 0.9960
• Recall Macro: 0.9778
• Precision Defect: 1.0
• Recall Defect: 0.9333
• F1 Defect: 0.9655
• Precision Empty: 1.0
• Recall Empty: 1.0
• F1 Empty: 1.0
• Precision Normal: 0.9880
• Recall Normal: 1.0
• F1 Normal: 0.9939

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: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 50

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1 Macro	Precision Macro	Recall Macro	Precision Defect	Recall Defect	F1 Defect	Precision Empty	Recall Empty	F1 Empty	Precision Normal	Recall Normal	F1 Normal
No log	1.0	4	0.9257	0.5490	0.3223	0.3463	0.3729	0.1818	0.5333	0.2712	0.0	0.0	0.0	0.8571	0.5854	0.6957
No log	2.0	8	0.4994	0.8137	0.4098	0.6040	0.4000	0.0	0.0	0.0	1.0	0.2	0.3333	0.8119	1.0	0.8962
0.8689	3.0	12	0.4794	0.8137	0.4098	0.6040	0.4000	0.0	0.0	0.0	1.0	0.2	0.3333	0.8119	1.0	0.8962
0.8689	4.0	16	0.3278	0.8725	0.7382	0.8903	0.6848	0.8	0.2667	0.4	1.0	0.8	0.8889	0.8710	0.9878	0.9257
0.3262	5.0	20	0.2778	0.8824	0.7551	0.9574	0.7333	1.0	0.2	0.3333	1.0	1.0	1.0	0.8723	1.0	0.9318
0.3262	6.0	24	0.1433	0.9510	0.9360	0.9293	0.9434	0.8125	0.8667	0.8387	1.0	1.0	1.0	0.9753	0.9634	0.9693
0.3262	7.0	28	0.1279	0.9706	0.9570	0.9882	0.9333	1.0	0.8	0.8889	1.0	1.0	1.0	0.9647	1.0	0.9820
0.0662	8.0	32	0.0992	0.9804	0.9722	0.9921	0.9556	1.0	0.8667	0.9286	1.0	1.0	1.0	0.9762	1.0	0.9880
0.0662	9.0	36	0.0342	0.9902	0.9865	0.9960	0.9778	1.0	0.9333	0.9655	1.0	1.0	1.0	0.9880	1.0	0.9939
0.0114	10.0	40	0.0251	0.9902	0.9865	0.9960	0.9778	1.0	0.9333	0.9655	1.0	1.0	1.0	0.9880	1.0	0.9939
0.0114	11.0	44	0.0439	0.9804	0.9722	0.9921	0.9556	1.0	0.8667	0.9286	1.0	1.0	1.0	0.9762	1.0	0.9880
0.0114	12.0	48	0.0327	0.9804	0.9722	0.9921	0.9556	1.0	0.8667	0.9286	1.0	1.0	1.0	0.9762	1.0	0.9880
0.0058	13.0	52	0.0291	0.9804	0.9722	0.9921	0.9556	1.0	0.8667	0.9286	1.0	1.0	1.0	0.9762	1.0	0.9880
0.0058	14.0	56	0.0266	0.9804	0.9722	0.9921	0.9556	1.0	0.8667	0.9286	1.0	1.0	1.0	0.9762	1.0	0.9880
0.0013	15.0	60	0.0255	0.9804	0.9722	0.9921	0.9556	1.0	0.8667	0.9286	1.0	1.0	1.0	0.9762	1.0	0.9880
0.0013	16.0	64	0.0277	0.9804	0.9722	0.9921	0.9556	1.0	0.8667	0.9286	1.0	1.0	1.0	0.9762	1.0	0.9880
0.0013	17.0	68	0.0312	0.9804	0.9722	0.9921	0.9556	1.0	0.8667	0.9286	1.0	1.0	1.0	0.9762	1.0	0.9880
0.0007	18.0	72	0.0351	0.9804	0.9722	0.9921	0.9556	1.0	0.8667	0.9286	1.0	1.0	1.0	0.9762	1.0	0.9880
0.0007	19.0	76	0.0384	0.9804	0.9722	0.9921	0.9556	1.0	0.8667	0.9286	1.0	1.0	1.0	0.9762	1.0	0.9880
0.0005	20.0	80	0.0402	0.9804	0.9722	0.9921	0.9556	1.0	0.8667	0.9286	1.0	1.0	1.0	0.9762	1.0	0.9880
0.0005	21.0	84	0.0409	0.9804	0.9722	0.9921	0.9556	1.0	0.8667	0.9286	1.0	1.0	1.0	0.9762	1.0	0.9880
0.0005	22.0	88	0.0408	0.9804	0.9722	0.9921	0.9556	1.0	0.8667	0.9286	1.0	1.0	1.0	0.9762	1.0	0.9880
0.0005	23.0	92	0.0412	0.9804	0.9722	0.9921	0.9556	1.0	0.8667	0.9286	1.0	1.0	1.0	0.9762	1.0	0.9880
0.0005	24.0	96	0.0414	0.9804	0.9722	0.9921	0.9556	1.0	0.8667	0.9286	1.0	1.0	1.0	0.9762	1.0	0.9880
0.0004	25.0	100	0.0412	0.9804	0.9722	0.9921	0.9556	1.0	0.8667	0.9286	1.0	1.0	1.0	0.9762	1.0	0.9880
0.0004	26.0	104	0.0413	0.9804	0.9722	0.9921	0.9556	1.0	0.8667	0.9286	1.0	1.0	1.0	0.9762	1.0	0.9880
0.0004	27.0	108	0.0415	0.9804	0.9722	0.9921	0.9556	1.0	0.8667	0.9286	1.0	1.0	1.0	0.9762	1.0	0.9880
0.0003	28.0	112	0.0419	0.9804	0.9722	0.9921	0.9556	1.0	0.8667	0.9286	1.0	1.0	1.0	0.9762	1.0	0.9880
0.0003	29.0	116	0.0431	0.9804	0.9722	0.9921	0.9556	1.0	0.8667	0.9286	1.0	1.0	1.0	0.9762	1.0	0.9880
0.0003	30.0	120	0.0458	0.9804	0.9722	0.9921	0.9556	1.0	0.8667	0.9286	1.0	1.0	1.0	0.9762	1.0	0.9880
0.0003	31.0	124	0.0474	0.9804	0.9722	0.9921	0.9556	1.0	0.8667	0.9286	1.0	1.0	1.0	0.9762	1.0	0.9880
0.0003	32.0	128	0.0484	0.9804	0.9722	0.9921	0.9556	1.0	0.8667	0.9286	1.0	1.0	1.0	0.9762	1.0	0.9880
0.0003	33.0	132	0.0489	0.9804	0.9722	0.9921	0.9556	1.0	0.8667	0.9286	1.0	1.0	1.0	0.9762	1.0	0.9880
0.0003	34.0	136	0.0492	0.9804	0.9722	0.9921	0.9556	1.0	0.8667	0.9286	1.0	1.0	1.0	0.9762	1.0	0.9880
0.0003	35.0	140	0.0492	0.9804	0.9722	0.9921	0.9556	1.0	0.8667	0.9286	1.0	1.0	1.0	0.9762	1.0	0.9880
0.0003	36.0	144	0.0491	0.9804	0.9722	0.9921	0.9556	1.0	0.8667	0.9286	1.0	1.0	1.0	0.9762	1.0	0.9880
0.0003	37.0	148	0.0490	0.9804	0.9722	0.9921	0.9556	1.0	0.8667	0.9286	1.0	1.0	1.0	0.9762	1.0	0.9880
0.0003	38.0	152	0.0489	0.9804	0.9722	0.9921	0.9556	1.0	0.8667	0.9286	1.0	1.0	1.0	0.9762	1.0	0.9880
0.0003	39.0	156	0.0485	0.9804	0.9722	0.9921	0.9556	1.0	0.8667	0.9286	1.0	1.0	1.0	0.9762	1.0	0.9880
0.0003	40.0	160	0.0481	0.9804	0.9722	0.9921	0.9556	1.0	0.8667	0.9286	1.0	1.0	1.0	0.9762	1.0	0.9880
0.0003	41.0	164	0.0478	0.9804	0.9722	0.9921	0.9556	1.0	0.8667	0.9286	1.0	1.0	1.0	0.9762	1.0	0.9880
0.0003	42.0	168	0.0475	0.9804	0.9722	0.9921	0.9556	1.0	0.8667	0.9286	1.0	1.0	1.0	0.9762	1.0	0.9880
0.0003	43.0	172	0.0473	0.9804	0.9722	0.9921	0.9556	1.0	0.8667	0.9286	1.0	1.0	1.0	0.9762	1.0	0.9880
0.0003	44.0	176	0.0471	0.9804	0.9722	0.9921	0.9556	1.0	0.8667	0.9286	1.0	1.0	1.0	0.9762	1.0	0.9880
0.0003	45.0	180	0.0468	0.9804	0.9722	0.9921	0.9556	1.0	0.8667	0.9286	1.0	1.0	1.0	0.9762	1.0	0.9880
0.0003	46.0	184	0.0467	0.9804	0.9722	0.9921	0.9556	1.0	0.8667	0.9286	1.0	1.0	1.0	0.9762	1.0	0.9880
0.0003	47.0	188	0.0466	0.9804	0.9722	0.9921	0.9556	1.0	0.8667	0.9286	1.0	1.0	1.0	0.9762	1.0	0.9880
0.0003	48.0	192	0.0466	0.9804	0.9722	0.9921	0.9556	1.0	0.8667	0.9286	1.0	1.0	1.0	0.9762	1.0	0.9880
0.0003	49.0	196	0.0465	0.9804	0.9722	0.9921	0.9556	1.0	0.8667	0.9286	1.0	1.0	1.0	0.9762	1.0	0.9880
0.0002	50.0	200	0.0465	0.9804	0.9722	0.9921	0.9556	1.0	0.8667	0.9286	1.0	1.0	1.0	0.9762	1.0	0.9880

Framework versions

• Transformers 4.56.1
• Pytorch 2.5.1+cu124
• Datasets 4.0.0
• Tokenizers 0.22.0