SWIN_finetuned_frozen_v5_cont

This model is a fine-tuned version of microsoft/swinv2-base-patch4-window12-192-22k on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 2.1793
• Accuracy: 0.7000

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: 0.0004
• train_batch_size: 256
• eval_batch_size: 256
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 15.0

Training results

Training Loss	Epoch	Step	Accuracy	Validation Loss
0.6306	1.0	2625	0.6565	1.9169
0.5977	2.0	5250	0.6614	1.9120
0.5472	3.0	7875	0.6635	1.9410
0.513	4.0	10500	0.6679	1.9818
0.4714	5.0	13125	0.6711	1.9428
0.4378	6.0	15750	0.6727	2.0051
0.4105	7.0	18375	0.6778	1.9917
0.3837	8.0	21000	0.6792	2.0413
0.3602	9.0	23625	0.6870	2.0738
0.336	10.0	26250	2.0872	0.6876
0.3057	11.0	28875	2.1163	0.6894
0.2856	12.0	31500	2.1105	0.6936
0.2704	13.0	34125	2.1685	0.6965
0.2503	14.0	36750	2.1627	0.6994
0.2362	15.0	39375	2.1793	0.7000

Framework versions

• Transformers 4.33.3
• Pytorch 2.1.2
• Datasets 2.16.1
• Tokenizers 0.13.3