swin-base-patch4-window7-224-in22k-construction_type

This model is a fine-tuned version of microsoft/swin-base-patch4-window7-224-in22k on the imagefolder dataset. It achieves the following results on the evaluation set:

• Loss: 0.3095
• Accuracy: 0.8804

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: 128
• eval_batch_size: 128
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 512
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 7

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
0.9839	0.9836	15	0.4599	0.8183
0.4167	1.9672	30	0.3605	0.8628
0.3853	2.9508	45	0.3272	0.8799
0.3302	4.0	61	0.3227	0.8763
0.3302	4.9836	76	0.3269	0.8753
0.3049	5.9672	91	0.3138	0.8799
0.2951	6.8852	105	0.3095	0.8804

Framework versions

• Transformers 4.44.2
• Pytorch 1.13.1+cu117
• Datasets 2.20.0
• Tokenizers 0.19.1