Shk4/vit_ana_0.89

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: 0.3768
• Accuracy: 0.8989

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: 64
• eval_batch_size: 64
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 256
• optimizer: Use 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
No log	0.9231	3	1.8889	0.2472
No log	1.8462	6	1.7625	0.3820
No log	2.7692	9	1.5603	0.4494
1.7854	4.0	13	1.3005	0.5281
1.7854	4.9231	16	1.0408	0.6292
1.7854	5.8462	19	0.8925	0.6854
1.1431	6.7692	22	0.7614	0.7303
1.1431	8.0	26	0.6343	0.7753
1.1431	8.9231	29	0.5810	0.7978
0.7715	9.8462	32	0.5551	0.8427
0.7715	10.7692	35	0.5209	0.8539
0.7715	12.0	39	0.5690	0.8202
0.5645	12.9231	42	0.4431	0.8876
0.5645	13.8462	45	0.4922	0.8202
0.5645	14.7692	48	0.4914	0.8315
0.4999	16.0	52	0.3768	0.8989
0.4999	16.9231	55	0.4292	0.8539
0.4999	17.8462	58	0.3846	0.8652
0.4555	18.7692	61	0.3498	0.8876
0.4555	20.0	65	0.3523	0.8652
0.4555	20.9231	68	0.3541	0.8876
0.3941	21.8462	71	0.3240	0.8989
0.3941	22.7692	74	0.3169	0.8989
0.3941	24.0	78	0.3317	0.8764
0.361	24.9231	81	0.3251	0.8876
0.361	25.8462	84	0.3198	0.8764
0.361	26.7692	87	0.3117	0.8764
0.3485	27.6923	90	0.3101	0.8764

Framework versions

• Transformers 4.46.3
• Pytorch 2.4.0
• Datasets 3.1.0
• Tokenizers 0.20.3