swinv2-tiny-patch4-window8-256-dmae-humeda-DAV6

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: 1.0899
• Accuracy: 0.5962

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: 64
• eval_batch_size: 64
• seed: 42
• 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: 30

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
1.65	1.0	12	1.5637	0.3462
1.5819	2.0	24	1.4962	0.3269
1.4527	3.0	36	1.3413	0.4615
1.2853	4.0	48	1.2034	0.4423
1.0185	5.0	60	1.1204	0.4808
0.8102	6.0	72	1.1004	0.4808
0.7477	7.0	84	1.1705	0.4808
0.6684	8.0	96	1.0245	0.5192
0.6297	9.0	108	1.0010	0.5577
0.5453	10.0	120	1.1905	0.4808
0.5496	11.0	132	1.1227	0.4808
0.4993	12.0	144	0.9619	0.5577
0.4297	13.0	156	1.0743	0.5192
0.4459	14.0	168	1.0195	0.5577
0.4219	15.0	180	1.0888	0.5
0.3742	16.0	192	1.0123	0.5769
0.4603	17.0	204	1.0503	0.5192
0.3607	18.0	216	1.1305	0.5577
0.3399	19.0	228	1.1327	0.5385
0.3422	20.0	240	1.1125	0.5192
0.3254	21.0	252	1.0243	0.5769
0.3363	22.0	264	1.0753	0.5577
0.3203	23.0	276	1.0778	0.5577
0.3248	24.0	288	1.1100	0.5385
0.2446	25.0	300	1.0773	0.5577
0.3058	26.0	312	1.0875	0.5769
0.254	27.0	324	1.0673	0.5769
0.2644	28.0	336	1.1026	0.5769
0.2962	29.0	348	1.0899	0.5962
0.2579	30.0	360	1.0816	0.5962

Framework versions

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