swinv2-tiny-patch4-window8-256-dmae-humeda-DAV49

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:

Model description

More information needed

More information needed

More information needed

The following hyperparameters were used during training:

learning_rate: 1.5e-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: cosine_with_restarts
lr_scheduler_warmup_ratio: 0.1
num_epochs: 40
mixed_precision_training: Native AMP

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	0.9412	8	1.5166	0.4318
1.5823	1.9412	16	1.4043	0.4432
1.5029	2.9412	24	1.3230	0.5
1.5029	3.9412	32	1.2373	0.5795
1.3569	4.9412	40	1.0701	0.6023
1.1064	5.9412	48	0.9832	0.6023
1.1064	6.9412	56	0.9004	0.6705
0.941	7.9412	64	0.8323	0.6591
0.7975	8.9412	72	0.7830	0.6818
0.7975	9.9412	80	0.7657	0.7045
0.7242	10.9412	88	0.7484	0.7386
0.6308	11.9412	96	0.7143	0.7386
0.6308	12.9412	104	0.6923	0.7273
0.5782	13.9412	112	0.6776	0.7386
0.5333	14.9412	120	0.6889	0.7614
0.5333	15.9412	128	0.6799	0.7841
0.495	16.9412	136	0.6794	0.7614
0.4931	17.9412	144	0.6921	0.7614
0.4931	18.9412	152	0.7162	0.7273
0.435	19.9412	160	0.7128	0.7386
0.4109	20.9412	168	0.7157	0.75
0.4109	21.9412	176	0.7404	0.7386
0.3897	22.9412	184	0.7275	0.7386
0.3718	23.9412	192	0.7492	0.7727
0.3718	24.9412	200	0.7520	0.7386
0.3866	25.9412	208	0.7550	0.7273
0.366	26.9412	216	0.7395	0.7386
0.366	27.9412	224	0.7340	0.7386
0.3454	28.9412	232	0.7578	0.7273
0.346	29.9412	240	0.7679	0.7273
0.346	30.9412	248	0.7546	0.75
0.3325	31.9412	256	0.7600	0.75
0.3117	32.9412	264	0.7798	0.7386
0.3117	33.9412	272	0.7944	0.7273
0.3177	34.9412	280	0.7856	0.7386
0.3263	35.9412	288	0.7813	0.7386
0.3263	36.9412	296	0.7798	0.7386
0.3305	37.9412	304	0.7804	0.7386
0.2999	38.9412	312	0.7810	0.7386
0.2999	39.9412	320	0.7811	0.7386