swin-tiny-patch4-window7-224-finetuned-piid

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: 0.5715
• Accuracy: 0.7854

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
1.2088	0.98	20	1.1661	0.4521
0.7545	2.0	41	0.8866	0.6073
0.6281	2.98	61	0.7788	0.6849
0.5939	4.0	82	0.6443	0.7397
0.5254	4.98	102	0.5097	0.7808
0.5583	6.0	123	0.5715	0.7854
0.3463	6.98	143	0.6163	0.7352
0.3878	8.0	164	0.5671	0.7671
0.3653	8.98	184	0.5690	0.7580
0.3529	10.0	205	0.5940	0.7580
0.301	10.98	225	0.6303	0.7626
0.2639	12.0	246	0.5725	0.7763
0.2847	12.98	266	0.6280	0.7717
0.25	14.0	287	0.5975	0.7717
0.2472	14.98	307	0.5821	0.7671
0.1676	16.0	328	0.6456	0.7626
0.1327	16.98	348	0.6117	0.7671
0.1977	18.0	369	0.6988	0.7489
0.1602	18.98	389	0.6448	0.7671
0.1785	19.51	400	0.6333	0.7717

Framework versions

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