swinv2-tiny-patch4-window8-256-dmae-humeda-DAV66

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: 0.2728
• Accuracy: 0.9086

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: 3e-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: 45
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
1.1391	1.0	15	1.0750	0.4057
0.9921	2.0	30	0.9329	0.6857
0.8345	3.0	45	0.5657	0.8229
0.5658	4.0	60	0.4445	0.8343
0.5066	5.0	75	0.4037	0.84
0.4016	6.0	90	0.3566	0.8629
0.4143	7.0	105	0.3347	0.8686
0.3599	8.0	120	0.3101	0.8743
0.3245	9.0	135	0.3187	0.88
0.3025	10.0	150	0.2728	0.9086
0.2546	11.0	165	0.2887	0.8743
0.2963	12.0	180	0.3797	0.8571
0.27	13.0	195	0.4297	0.84
0.2374	14.0	210	0.4063	0.8457
0.2626	15.0	225	0.3277	0.8914
0.2268	16.0	240	0.3114	0.8914
0.1787	17.0	255	0.3716	0.8571
0.2187	18.0	270	0.3290	0.88
0.1818	19.0	285	0.3367	0.9029
0.1822	20.0	300	0.3245	0.8857
0.1747	21.0	315	0.3293	0.88
0.1728	22.0	330	0.3330	0.88
0.1688	23.0	345	0.3699	0.8857
0.1953	24.0	360	0.3193	0.9029
0.1677	25.0	375	0.3689	0.8914
0.1658	26.0	390	0.3698	0.8971
0.147	27.0	405	0.3759	0.8914
0.1599	28.0	420	0.4673	0.8571
0.1165	29.0	435	0.4239	0.88
0.1235	30.0	450	0.4119	0.8857
0.098	31.0	465	0.4140	0.8914
0.1042	32.0	480	0.4382	0.88
0.1052	33.0	495	0.4593	0.8743
0.1021	34.0	510	0.5186	0.8629
0.1074	35.0	525	0.4545	0.88
0.0964	36.0	540	0.4612	0.88
0.1047	37.0	555	0.4607	0.8743
0.1218	38.0	570	0.4805	0.8629
0.1073	39.0	585	0.4456	0.8914
0.0948	40.0	600	0.4985	0.8743
0.1087	41.0	615	0.4687	0.8857
0.1055	42.0	630	0.4863	0.8857
0.1082	43.0	645	0.4663	0.8914
0.0829	44.0	660	0.4579	0.8857
0.0774	45.0	675	0.4660	0.8857

Framework versions

• Transformers 4.49.0
• Pytorch 2.6.0+cu124
• Datasets 3.4.1
• Tokenizers 0.21.1