swinv2-large-patch4-window12-192-22k-finetuned-ethzurich

This model is a fine-tuned version of microsoft/swinv2-large-patch4-window12-192-22k on the Urban Resource Cadastre dataset created by Deepika Raghu, Martin Juan José Bucher, and Catherine De Wolf (https://github.com/raghudeepika/urban-resource-cadastre-repository). It achieves the following results on the evaluation set:

• Loss: 0.6083
• Accuracy: 0.8295

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: 3

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	0.96	6	1.2578	0.6364
1.6142	1.92	12	0.7696	0.75
1.6142	2.88	18	0.6083	0.8295

Framework versions

• Transformers 4.33.3
• Pytorch 2.0.1+cu117
• Datasets 2.14.5
• Tokenizers 0.13.3