swin-tiny-patch4-window7-224-MM_Classification

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: 0.3468
• Accuracy: 0.8694

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: 128
• eval_batch_size: 128
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 512
• 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
1.0476	1.0	19	0.7707	0.6530
0.6226	2.0	38	0.4743	0.8105
0.4477	3.0	57	0.4133	0.8323
0.3963	4.0	76	0.3813	0.8476
0.3694	5.0	95	0.3753	0.8540
0.3451	6.0	114	0.3587	0.8489
0.3382	7.0	133	0.3531	0.8451
0.3253	8.0	152	0.3498	0.8579
0.3121	9.0	171	0.3437	0.8579
0.2855	10.0	190	0.3447	0.8656
0.2961	11.0	209	0.3350	0.8617
0.273	12.0	228	0.3484	0.8566
0.2745	13.0	247	0.3433	0.8604
0.2613	14.0	266	0.3498	0.8643
0.2527	15.0	285	0.3365	0.8579
0.2619	16.0	304	0.3450	0.8617
0.2436	17.0	323	0.3454	0.8681
0.2518	18.0	342	0.3437	0.8681
0.243	19.0	361	0.3468	0.8694
0.2415	20.0	380	0.3455	0.8694

Framework versions

• Transformers 4.43.3
• Pytorch 1.13.1+cu117
• Datasets 2.20.0
• Tokenizers 0.19.1