Boya2_SGD_1e3_20Epoch_Swin-large-224_fold2

This model is a fine-tuned version of microsoft/swin-large-patch4-window7-224-in22k on the imagefolder dataset. It achieves the following results on the evaluation set:

• Loss: 1.7316
• Accuracy: 0.4624

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: 0.001
• train_batch_size: 16
• eval_batch_size: 16
• seed: 42
• 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
2.613	1.0	913	2.5089	0.2774
2.2378	2.0	1826	2.3352	0.2938
2.1303	3.0	2739	2.2226	0.3184
2.0563	4.0	3652	2.1385	0.3506
2.1322	5.0	4565	2.0717	0.3758
2.004	6.0	5478	2.0113	0.3955
2.079	7.0	6391	1.9666	0.4130
1.78	8.0	7304	1.9230	0.4225
1.7925	9.0	8217	1.8877	0.4354
1.9366	10.0	9130	1.8567	0.4403
1.5845	11.0	10043	1.8338	0.4458
1.6846	12.0	10956	1.8079	0.4520
1.7316	13.0	11869	1.7920	0.4545
1.6397	14.0	12782	1.7737	0.4600
1.8114	15.0	13695	1.7580	0.4605
1.5862	16.0	14608	1.7502	0.4621
1.8464	17.0	15521	1.7422	0.4641
1.8008	18.0	16434	1.7370	0.4627
1.719	19.0	17347	1.7333	0.4632
1.5652	20.0	18260	1.7316	0.4624

Framework versions

• Transformers 4.32.1
• Pytorch 2.1.1+cu121
• Datasets 2.21.0
• Tokenizers 0.13.2