Electrcical-IMAGE-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: 0.3583
• Accuracy: 0.8960

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
1.6143	0.98	28	1.2882	0.5347
0.8597	2.0	57	0.7302	0.7649
0.5858	2.98	85	0.4849	0.8465
0.4332	4.0	114	0.4274	0.8614
0.4054	4.98	142	0.3687	0.8787
0.3826	6.0	171	0.3788	0.8614
0.3561	6.98	199	0.3700	0.8936
0.2838	8.0	228	0.3550	0.8812
0.2897	8.98	256	0.3698	0.8886
0.2519	10.0	285	0.3459	0.8837
0.2194	10.98	313	0.3583	0.8960
0.1955	12.0	342	0.3442	0.8886
0.2443	12.98	370	0.3801	0.8787
0.207	14.0	399	0.3499	0.8861
0.2078	14.98	427	0.3701	0.8837
0.1873	16.0	456	0.3773	0.8861
0.1697	16.98	484	0.3753	0.8861
0.1812	18.0	513	0.3747	0.8911
0.151	18.98	541	0.3736	0.8861
0.1567	19.65	560	0.3726	0.8861

Framework versions

• Transformers 4.39.3
• Pytorch 2.1.2
• Datasets 2.18.0
• Tokenizers 0.15.2