swin-small-patch4-window7-224-finetuned-piid

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

• Loss: 0.6168
• Accuracy: 0.7763

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
1.2327	0.98	20	1.1687	0.5114
0.7354	2.0	41	0.7696	0.6712
0.602	2.98	61	0.7198	0.7078
0.5809	4.0	82	0.5824	0.7397
0.4989	4.98	102	0.5331	0.7489
0.4364	6.0	123	0.6137	0.7489
0.3321	6.98	143	0.5839	0.7717
0.3	8.0	164	0.5246	0.7763
0.3024	8.98	184	0.5557	0.7717
0.3433	10.0	205	0.5258	0.7900
0.258	10.98	225	0.6354	0.7489
0.1595	12.0	246	0.5492	0.8219
0.2295	12.98	266	0.5889	0.7900
0.1956	14.0	287	0.5670	0.7900
0.2028	14.98	307	0.5460	0.7900
0.1514	16.0	328	0.6587	0.7900
0.0934	16.98	348	0.6131	0.7945
0.1323	18.0	369	0.6615	0.7900
0.1213	18.98	389	0.6192	0.7671
0.1028	19.51	400	0.6168	0.7763

Framework versions

• Transformers 4.35.0
• Pytorch 2.1.0+cu118
• Datasets 2.14.6
• Tokenizers 0.14.1