Edit model card

swin-tiny-patch4-window7-224-finetuned-omars6

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: 0.5625
  • Accuracy: 0.8815

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: 0.0005
  • 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: 30

Training results

Training Loss Epoch Step Validation Loss Accuracy
0.9598 0.99 92 0.7744 0.6869
0.7825 2.0 185 0.7336 0.7082
0.9638 2.99 277 0.8202 0.7204
1.0288 4.0 370 0.8621 0.7903
0.9711 4.99 462 0.8212 0.6809
1.0125 6.0 555 0.8700 0.7356
0.945 6.99 647 0.7959 0.7781
0.9851 8.0 740 0.8755 0.6140
0.8078 8.99 832 0.6970 0.7781
0.7377 10.0 925 0.6063 0.7386
0.7934 10.99 1017 0.6121 0.8116
0.7986 12.0 1110 0.6532 0.8116
0.6129 12.99 1202 0.7250 0.8450
0.7428 14.0 1295 0.6417 0.7264
0.5661 14.99 1387 0.6847 0.7964
0.6631 16.0 1480 0.5470 0.8298
0.5787 16.99 1572 0.5696 0.8359
0.6635 18.0 1665 0.6385 0.7872
0.5251 18.99 1757 0.5842 0.8419
0.6164 20.0 1850 0.5506 0.8207
0.4166 20.99 1942 0.8169 0.8055
0.4189 22.0 2035 0.5882 0.8480
0.699 22.99 2127 0.5767 0.8541
0.6095 24.0 2220 0.6392 0.8845
0.3837 24.99 2312 0.6109 0.8723
0.4916 26.0 2405 0.4862 0.8754
0.4536 26.99 2497 0.5625 0.8754
0.3636 28.0 2590 0.5948 0.8663
0.4004 28.99 2682 0.5735 0.8906
0.4248 29.84 2760 0.5625 0.8815

Framework versions

  • Transformers 4.30.2
  • Pytorch 2.0.1+cu117
  • Datasets 2.13.0
  • Tokenizers 0.13.3
Downloads last month
11
Inference Examples
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social visibility and check back later, or deploy to Inference Endpoints (dedicated) instead.

Evaluation results