swinv2-tiny-patch4-window8-256-dmae-humeda-DAV19

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: 1.4438
• Accuracy: 0.7308

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: 7e-05
• train_batch_size: 64
• eval_batch_size: 64
• seed: 42
• 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: 0.3
• num_epochs: 42

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
1.6103	1.0	22	1.6051	0.2885
1.4727	2.0	44	1.4581	0.4808
1.0885	3.0	66	1.1148	0.5385
0.8317	4.0	88	1.1753	0.4808
0.5767	5.0	110	1.0842	0.5192
0.4983	6.0	132	0.9595	0.5769
0.4392	7.0	154	0.9108	0.6538
0.38	8.0	176	0.7973	0.6923
0.367	9.0	198	0.8772	0.6346
0.2892	10.0	220	0.9240	0.6346
0.2627	11.0	242	1.1102	0.6154
0.1956	12.0	264	0.8497	0.7115
0.2529	13.0	286	0.9588	0.6923
0.1933	14.0	308	1.4496	0.5962
0.2023	15.0	330	1.2467	0.6346
0.1725	16.0	352	1.1693	0.6731
0.1604	17.0	374	1.1374	0.6346
0.1909	18.0	396	0.9065	0.7115
0.1577	19.0	418	1.1488	0.6538
0.1323	20.0	440	1.3994	0.6923
0.1342	21.0	462	1.1350	0.6731
0.1024	22.0	484	1.2422	0.6538
0.1054	23.0	506	1.0670	0.75
0.0809	24.0	528	1.2367	0.6731
0.0856	25.0	550	1.1758	0.7308
0.0781	26.0	572	1.1735	0.6731
0.1136	27.0	594	1.5008	0.6923
0.0784	28.0	616	1.2966	0.7308
0.0648	29.0	638	1.2018	0.7115
0.0941	30.0	660	1.0879	0.6731
0.0654	31.0	682	1.2646	0.7115
0.0967	32.0	704	1.0537	0.75
0.0717	33.0	726	1.4332	0.7115
0.0715	34.0	748	1.2683	0.7308
0.0773	35.0	770	1.3363	0.6731
0.0767	36.0	792	1.3192	0.6731
0.0343	37.0	814	1.2926	0.7115
0.0524	38.0	836	1.4072	0.7115
0.052	39.0	858	1.4377	0.6923
0.0247	40.0	880	1.4420	0.6923
0.0256	41.0	902	1.4403	0.7115
0.0384	42.0	924	1.4438	0.7308

Framework versions

• Transformers 4.47.1
• Pytorch 2.5.1+cu121
• Datasets 3.2.0
• Tokenizers 0.21.0