HITMYM/practice_swin1

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: 0.5657
• Accuracy: 0.7821

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
2.1465	0.97	16	1.8341	0.3462
1.7722	2.0	33	1.5865	0.4017
1.6005	2.97	49	1.4867	0.4060
1.429	4.0	66	1.3933	0.4487
1.2294	4.97	82	1.2696	0.5385
1.1224	6.0	99	1.2842	0.5641
0.9776	6.97	115	0.9923	0.6197
0.8678	8.0	132	1.1118	0.6368
0.8125	8.97	148	0.8974	0.6624
0.7022	10.0	165	0.8582	0.6838
0.6047	10.97	181	0.7019	0.7393
0.6223	12.0	198	0.6818	0.7308
0.5331	12.97	214	0.8265	0.7051
0.4995	14.0	231	0.6365	0.7521
0.4132	14.97	247	0.6585	0.7308
0.3978	16.0	264	0.5789	0.7692
0.3388	16.97	280	0.6038	0.7650
0.3376	18.0	297	0.5306	0.7821
0.3455	18.97	313	0.5797	0.7692
0.3207	19.39	320	0.5657	0.7821

Framework versions

• Transformers 4.30.2
• Pytorch 2.0.1+cu118
• Datasets 2.13.0
• Tokenizers 0.13.3