swinv2-tiny-patch4-window8-256-dmae-humeda-DAV48

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: 0.7283
• Accuracy: 0.75

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: 2e-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: reduce_lr_on_plateau
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 40
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	0.9412	8	1.4387	0.4432
1.5789	1.9412	16	1.3131	0.5568
1.3907	2.9412	24	1.1805	0.5909
1.3907	3.9412	32	1.0386	0.6136
1.1967	4.9412	40	1.0065	0.6136
1.0098	5.9412	48	0.8786	0.6477
1.0098	6.9412	56	0.8264	0.6932
0.863	7.9412	64	0.8026	0.7273
0.7309	8.9412	72	0.7853	0.7159
0.7309	9.9412	80	0.7649	0.7273
0.6597	10.9412	88	0.7671	0.7386
0.56	11.9412	96	0.7551	0.7159
0.56	12.9412	104	0.7428	0.7273
0.5207	13.9412	112	0.7396	0.7273
0.5108	14.9412	120	0.7368	0.7273
0.5108	15.9412	128	0.7366	0.7386
0.5062	16.9412	136	0.7364	0.7273
0.5069	17.9412	144	0.7329	0.7386
0.5069	18.9412	152	0.7285	0.7273
0.4952	19.9412	160	0.7371	0.7386
0.4979	20.9412	168	0.7436	0.7386
0.4979	21.9412	176	0.7338	0.7386
0.4745	22.9412	184	0.7291	0.75
0.4735	23.9412	192	0.7305	0.75
0.4735	24.9412	200	0.7301	0.75
0.4862	25.9412	208	0.7283	0.75
0.4955	26.9412	216	0.7273	0.75
0.4955	27.9412	224	0.7275	0.75
0.4602	28.9412	232	0.7280	0.75
0.4714	29.9412	240	0.7291	0.75
0.4714	30.9412	248	0.7298	0.75
0.4727	31.9412	256	0.7301	0.75
0.4689	32.9412	264	0.7293	0.75
0.4689	33.9412	272	0.7287	0.75
0.4725	34.9412	280	0.7287	0.75
0.4747	35.9412	288	0.7284	0.75
0.4747	36.9412	296	0.7284	0.75
0.5012	37.9412	304	0.7284	0.75
0.462	38.9412	312	0.7286	0.75
0.462	39.9412	320	0.7283	0.75

Framework versions

• Transformers 4.48.2
• Pytorch 2.5.1+cu124
• Datasets 3.2.0
• Tokenizers 0.21.0