swinv2-tiny-patch4-window8-256-dmae-humeda-DAV63

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: 0.2824
• 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: 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: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 45
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	1.0	5	1.0434	0.4743
1.0715	2.0	10	1.0274	0.4571
1.0715	3.0	15	0.9232	0.5714
0.9224	4.0	20	0.7760	0.7257
0.9224	5.0	25	0.6111	0.7314
0.659	6.0	30	0.4303	0.8514
0.659	7.0	35	0.3456	0.8743
0.4725	8.0	40	0.4387	0.8
0.4725	9.0	45	0.3622	0.8286
0.4171	10.0	50	0.3593	0.8686
0.4171	11.0	55	0.3262	0.8629
0.3615	12.0	60	0.3256	0.8571
0.3615	13.0	65	0.3205	0.8743
0.3094	14.0	70	0.3051	0.8743
0.3094	15.0	75	0.2948	0.8857
0.2784	16.0	80	0.3237	0.8857
0.2784	17.0	85	0.2920	0.88
0.2849	18.0	90	0.2824	0.9029
0.2849	19.0	95	0.3579	0.8743
0.2772	20.0	100	0.2966	0.8857
0.2772	21.0	105	0.3484	0.8743
0.2135	22.0	110	0.3230	0.88
0.2135	23.0	115	0.3591	0.8686
0.2067	24.0	120	0.3292	0.8743
0.2067	25.0	125	0.3496	0.88
0.165	26.0	130	0.3628	0.88
0.165	27.0	135	0.3345	0.8914
0.1888	28.0	140	0.3596	0.88
0.1888	29.0	145	0.3567	0.8914
0.1473	30.0	150	0.3824	0.8743
0.1473	31.0	155	0.3685	0.88
0.1472	32.0	160	0.3488	0.8971
0.1472	33.0	165	0.3895	0.8857
0.1568	34.0	170	0.3929	0.8914
0.1568	35.0	175	0.3803	0.8857
0.129	36.0	180	0.3954	0.8743
0.129	37.0	185	0.3920	0.8743
0.1295	38.0	190	0.3945	0.8971
0.1295	39.0	195	0.3836	0.8857
0.1322	40.0	200	0.4102	0.8743
0.1322	41.0	205	0.4226	0.8686
0.1124	42.0	210	0.3939	0.88
0.1124	43.0	215	0.3911	0.8857
0.1162	44.0	220	0.3986	0.8914
0.1162	45.0	225	0.4015	0.8857

Framework versions

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