swinv2-tiny-patch4-window8-256-dmae-humeda-DAV65

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: 0.4432
• Accuracy: 0.9029

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: 6e-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
1.1096	1.0	15	1.0206	0.48
0.7949	2.0	30	0.5637	0.8057
0.6305	3.0	45	0.3961	0.8457
0.4475	4.0	60	0.3464	0.8629
0.4342	5.0	75	0.3489	0.8629
0.3843	6.0	90	0.3304	0.8457
0.3059	7.0	105	0.3037	0.88
0.2993	8.0	120	0.3286	0.8686
0.2799	9.0	135	0.4499	0.8457
0.302	10.0	150	0.2965	0.8857
0.2445	11.0	165	0.2569	0.8857
0.2317	12.0	180	0.4269	0.8286
0.2189	13.0	195	0.4250	0.8514
0.2001	14.0	210	0.3725	0.8743
0.1924	15.0	225	0.3042	0.8571
0.1617	16.0	240	0.4525	0.8686
0.131	17.0	255	0.4021	0.8457
0.1491	18.0	270	0.3316	0.8571
0.1576	19.0	285	0.3288	0.8857
0.1189	20.0	300	0.3464	0.88
0.1384	21.0	315	0.3618	0.8857
0.1109	22.0	330	0.3548	0.8914
0.0985	23.0	345	0.3908	0.8857
0.1095	24.0	360	0.4517	0.8743
0.1007	25.0	375	0.5406	0.8743
0.192	26.0	390	0.5257	0.8686
0.093	27.0	405	0.4442	0.8686
0.1337	28.0	420	0.5376	0.8629
0.0737	29.0	435	0.4627	0.8686
0.0932	30.0	450	0.4371	0.8914
0.0638	31.0	465	0.4741	0.8971
0.0796	32.0	480	0.4220	0.88
0.0674	33.0	495	0.4432	0.9029
0.0466	34.0	510	0.4385	0.8914
0.0586	35.0	525	0.4614	0.8971
0.0634	36.0	540	0.4855	0.8857
0.0867	37.0	555	0.4716	0.88
0.0721	38.0	570	0.4353	0.8914
0.0572	39.0	585	0.4443	0.8914
0.0613	40.0	600	0.4655	0.8857
0.0911	41.0	615	0.4462	0.8971
0.0691	42.0	630	0.4922	0.88
0.0767	43.0	645	0.4702	0.8743
0.0763	44.0	660	0.4670	0.8743
0.055	45.0	675	0.4692	0.8686

Framework versions

• Transformers 4.48.3
• Pytorch 2.6.0+cu124
• Datasets 3.4.1
• Tokenizers 0.21.1