swinv2-tiny-patch4-window8-256-dmae-humeda-DAV45

This model is a fine-tuned version of microsoft/swinv2-tiny-patch4-window8-256 on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.7812
• Accuracy: 0.7841

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: 32
• eval_batch_size: 32
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 128
• optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: cosine
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 42

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
6.1945	1.0	20	1.2588	0.4545
4.5836	2.0	40	0.9658	0.7159
2.9056	3.0	60	0.7737	0.7273
2.8061	4.0	80	0.6738	0.7727
1.9405	5.0	100	0.6261	0.7614
1.4425	6.0	120	0.8127	0.75
1.3554	7.0	140	0.7812	0.7841
1.2975	8.0	160	0.8405	0.75
0.812	9.0	180	1.0777	0.7159
0.7984	10.0	200	0.9404	0.7159
0.7895	11.0	220	1.0902	0.7045
0.7333	12.0	240	1.0998	0.75
0.6073	13.0	260	1.2734	0.7386
0.6548	14.0	280	1.3034	0.7159
0.5538	15.0	300	1.1890	0.75
0.556	16.0	320	1.3662	0.75
0.5273	17.0	340	1.2833	0.7273
0.3863	18.0	360	1.2976	0.7159
0.5185	19.0	380	1.2461	0.7386
0.475	20.0	400	1.2543	0.7386
0.3021	21.0	420	1.3143	0.7727
0.3334	22.0	440	1.2873	0.75
0.3773	23.0	460	1.3992	0.7386
0.2606	24.0	480	1.5181	0.7159
0.3344	25.0	500	1.4330	0.7614
0.3349	26.0	520	1.4165	0.7841
0.3246	27.0	540	1.3634	0.7614
0.3395	28.0	560	1.3985	0.7614
0.2606	29.0	580	1.3866	0.7614
0.2212	30.0	600	1.4849	0.75
0.2266	31.0	620	1.4230	0.7727
0.2525	32.0	640	1.4288	0.7727
0.2241	33.0	660	1.4497	0.7614
0.1816	34.0	680	1.4347	0.7614
0.2529	35.0	700	1.4278	0.75
0.189	36.0	720	1.4290	0.75
0.2491	37.0	740	1.4449	0.7614
0.2562	38.0	760	1.4514	0.75
0.1872	39.0	780	1.4522	0.75
0.223	39.9351	798	1.4527	0.75

Framework versions

• Transformers 4.47.1
• Pytorch 2.5.1+cu124
• Datasets 3.2.0
• Tokenizers 0.21.0