swinv2-tiny-patch4-window8-256-finetuned-eurosat

This model is a fine-tuned version of microsoft/swinv2-tiny-patch4-window8-256 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.6760
• Accuracy: 0.8170

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: 8
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 32
• 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
0.2519	0.9955	167	0.7302	0.8017
0.2407	1.9970	335	0.7095	0.7836
0.3423	2.9985	503	0.7016	0.7884
0.4687	4.0	671	0.6480	0.7969
0.4789	4.9955	838	0.5132	0.8160
0.4417	5.9970	1006	0.5321	0.8065
0.435	6.9985	1174	0.5770	0.8093
0.4106	8.0	1342	0.5650	0.8189
0.4216	8.9955	1509	0.5535	0.8132
0.3786	9.9970	1677	0.5745	0.8179
0.3536	10.9985	1845	0.6322	0.8046
0.4842	12.0	2013	0.7200	0.8103
0.3095	12.9955	2180	0.6996	0.8112
0.2603	13.9970	2348	0.7004	0.8065
0.2838	14.9985	2516	0.6331	0.8227
0.3449	16.0	2684	0.6788	0.8122
0.253	16.9955	2851	0.6940	0.8103
0.2647	17.9970	3019	0.6770	0.8132
0.2991	18.9985	3187	0.6647	0.8189
0.26	19.9106	3340	0.6760	0.8170

Framework versions

• Transformers 4.40.2
• Pytorch 2.2.2
• Datasets 2.19.1
• Tokenizers 0.19.1