swinv2-tiny-patch4-window8-256-dmae-humeda-DAV31

This model is a fine-tuned version of microsoft/swinv2-tiny-patch4-window8-256 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 3.0993
• Accuracy: 0.5902

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: 5e-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: 40

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	1.0	5	1.5163	0.4590
No log	2.0	10	1.3238	0.4590
5.237	3.0	15	1.2645	0.5082
5.237	4.0	20	1.1982	0.5082
3.9571	5.0	25	1.1256	0.5246
3.9571	6.0	30	1.2419	0.5902
3.9571	7.0	35	1.1806	0.4918
2.5954	8.0	40	1.3698	0.5738
2.5954	9.0	45	1.3558	0.5574
1.7082	10.0	50	1.5148	0.5902
1.7082	11.0	55	1.6398	0.5574
0.7716	12.0	60	1.6954	0.6066
0.7716	13.0	65	1.8204	0.6066
0.7716	14.0	70	1.8661	0.6066
0.3952	15.0	75	1.9419	0.6066
0.3952	16.0	80	2.1287	0.5902
0.144	17.0	85	2.4729	0.5738
0.144	18.0	90	2.8243	0.5410
0.144	19.0	95	2.6554	0.5902
0.0821	20.0	100	2.5922	0.5574
0.0821	21.0	105	2.9639	0.4918
0.0948	22.0	110	2.8189	0.5738
0.0948	23.0	115	2.8026	0.5410
0.0809	24.0	120	2.6857	0.6066
0.0809	25.0	125	2.8069	0.5574
0.0809	26.0	130	2.7649	0.5902
0.0309	27.0	135	2.8005	0.5902
0.0309	28.0	140	2.8427	0.6066
0.0269	29.0	145	2.9175	0.6066
0.0269	30.0	150	3.0034	0.5738
0.0269	31.0	155	3.0874	0.5902
0.0279	32.0	160	3.0993	0.5902

Framework versions

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