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: 0.4435
• Accuracy: 0.8111

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: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
0.5077	0.98	41	0.6378	0.6796
0.5111	1.99	83	0.7097	0.6577
0.5395	2.99	125	0.5374	0.7470
0.5498	4.0	167	0.5524	0.7420
0.4754	4.98	208	0.5324	0.7639
0.4662	5.99	250	0.4962	0.7639
0.4677	6.99	292	0.5070	0.7774
0.4525	8.0	334	0.5144	0.7673
0.4635	8.98	375	0.4978	0.7757
0.4309	9.99	417	0.5388	0.7774
0.4292	10.99	459	0.4937	0.7825
0.4182	12.0	501	0.5234	0.7808
0.4242	12.98	542	0.4539	0.7960
0.4053	13.99	584	0.5089	0.7858
0.4135	14.99	626	0.4655	0.8044
0.3888	16.0	668	0.4398	0.8212
0.3701	16.98	709	0.4258	0.8145
0.3641	17.99	751	0.4339	0.8196
0.3547	18.99	793	0.4556	0.7993
0.3623	19.64	820	0.4435	0.8111

Framework versions

• Transformers 4.35.2
• Pytorch 2.1.0+cu118
• Datasets 2.15.0
• Tokenizers 0.15.0