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: 0.6247
• Accuracy: 0.8279

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: 30

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
1.6684	1.0	43	1.5274	0.5386
1.422	2.0	86	1.2167	0.5964
1.2303	3.0	129	0.9840	0.6617
1.0419	4.0	172	0.8583	0.7255
0.886	5.0	215	0.7485	0.7389
0.8443	6.0	258	0.7898	0.7463
0.7685	7.0	301	0.6678	0.7745
0.6999	8.0	344	0.7002	0.7730
0.6128	9.0	387	0.6634	0.7982
0.5588	10.0	430	0.6644	0.7789
0.5829	11.0	473	0.6318	0.8116
0.5234	12.0	516	0.6662	0.7864
0.4712	13.0	559	0.6781	0.8042
0.4042	14.0	602	0.6542	0.8131
0.3966	15.0	645	0.6432	0.8086
0.41	16.0	688	0.6346	0.8145
0.3848	17.0	731	0.6295	0.8323
0.3612	18.0	774	0.6841	0.8042
0.3258	19.0	817	0.6613	0.8145
0.3163	20.0	860	0.6340	0.8279
0.3469	21.0	903	0.6621	0.8205
0.3523	22.0	946	0.6655	0.8131
0.3533	23.0	989	0.6541	0.8131
0.3312	24.0	1032	0.6445	0.8116
0.3095	25.0	1075	0.6519	0.8205
0.2425	26.0	1118	0.6363	0.8145
0.2956	27.0	1161	0.6318	0.8294
0.2629	28.0	1204	0.6217	0.8249
0.2755	29.0	1247	0.6243	0.8279
0.29	29.3077	1260	0.6247	0.8279

Framework versions

• Transformers 4.50.0
• Pytorch 2.6.0+cu124
• Datasets 3.5.0
• Tokenizers 0.21.1