swin-tiny-patch4-window7-224-finetuned_ASL_Isolated_Swin_dataset2

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

• Loss: 0.1269
• Accuracy: 0.9769

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: 0.0002
• train_batch_size: 16
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
1.5439	1.09	100	1.4188	0.5538
0.8646	2.17	200	0.4542	0.8885
0.5485	3.26	300	0.4103	0.8538
0.5082	4.35	400	0.2925	0.8962
0.5302	5.43	500	0.2471	0.9269
0.4072	6.52	600	0.2676	0.9231
0.4424	7.61	700	0.4150	0.9038
0.3409	8.7	800	0.1922	0.9538
0.3046	9.78	900	0.1917	0.9462
0.2911	10.87	1000	0.2272	0.9423
0.269	11.96	1100	0.0722	0.9692
0.3709	13.04	1200	0.1473	0.9654
0.3443	14.13	1300	0.1545	0.9615
0.187	15.22	1400	0.1060	0.9731
0.1879	16.3	1500	0.1124	0.9692
0.2183	17.39	1600	0.1377	0.9615
0.1478	18.48	1700	0.1269	0.9769
0.1944	19.57	1800	0.0909	0.9769

Framework versions

• Transformers 4.34.1
• Pytorch 2.1.0+cu118
• Datasets 2.14.6
• Tokenizers 0.14.1