swinv2-tiny-patch4-window8-256-dmae-humeda-DAV73

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: 0.3467
• Accuracy: 0.8914

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: 4e-05
• train_batch_size: 16
• eval_batch_size: 16
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 64
• 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: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 50
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
1.1143	1.0	15	1.0550	0.5886
0.9057	2.0	30	0.7895	0.7314
0.7128	3.0	45	0.5484	0.76
0.491	4.0	60	0.4717	0.7829
0.4667	5.0	75	0.3817	0.8457
0.39	6.0	90	0.3647	0.88
0.3975	7.0	105	0.3604	0.8571
0.3496	8.0	120	0.3397	0.8457
0.3546	9.0	135	0.3968	0.84
0.2798	10.0	150	0.3983	0.8571
0.2813	11.0	165	0.3591	0.8571
0.2178	12.0	180	0.4332	0.8571
0.234	13.0	195	0.3892	0.8686
0.2539	14.0	210	0.3944	0.8743
0.2147	15.0	225	0.5366	0.8286
0.2045	16.0	240	0.3467	0.8914
0.1826	17.0	255	0.4077	0.8629
0.1703	18.0	270	0.4043	0.8743
0.1928	19.0	285	0.4470	0.88
0.1853	20.0	300	0.4742	0.8857
0.1503	21.0	315	0.5047	0.8743
0.1692	22.0	330	0.4166	0.8686
0.1436	23.0	345	0.4770	0.8686
0.1715	24.0	360	0.4240	0.8514
0.1449	25.0	375	0.4297	0.8743
0.1051	26.0	390	0.5423	0.8743
0.1055	27.0	405	0.5705	0.8743
0.1407	28.0	420	0.5883	0.8743
0.1463	29.0	435	0.4944	0.8857
0.1413	30.0	450	0.5196	0.8914
0.1521	31.0	465	0.6213	0.8571
0.1318	32.0	480	0.7071	0.84
0.1357	33.0	495	0.5737	0.8743
0.1182	34.0	510	0.5457	0.88
0.1382	35.0	525	0.5622	0.8743
0.1108	36.0	540	0.5403	0.88
0.0948	37.0	555	0.6038	0.8629
0.0964	38.0	570	0.6299	0.8743
0.1035	39.0	585	0.5474	0.8857
0.0892	40.0	600	0.5423	0.8914
0.1121	41.0	615	0.5783	0.8686
0.1119	42.0	630	0.6223	0.8743
0.1103	43.0	645	0.6355	0.88
0.0934	44.0	660	0.5847	0.88
0.0861	45.0	675	0.6008	0.88
0.0933	46.0	690	0.6066	0.88
0.0909	46.6780	700	0.6084	0.88

Framework versions

• Transformers 4.51.3
• Pytorch 2.6.0+cu124
• Datasets 2.19.0
• Tokenizers 0.21.1