swinv2-tiny-patch4-window8-256-dmae-humeda-DAV74

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: 0.3840
• Accuracy: 0.9314

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: 4e-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: 100
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
1.1566	1.0	15	1.1332	0.2286
1.0369	2.0	30	0.9972	0.6171
0.9264	3.0	45	0.6970	0.7714
0.629	4.0	60	0.4858	0.8286
0.5428	5.0	75	0.4445	0.8171
0.4508	6.0	90	0.3967	0.8457
0.4188	7.0	105	0.3711	0.8286
0.3688	8.0	120	0.3448	0.8629
0.3503	9.0	135	0.3299	0.8743
0.2885	10.0	150	0.2780	0.8971
0.2582	11.0	165	0.3197	0.8686
0.3041	12.0	180	0.4322	0.8514
0.2957	13.0	195	0.4167	0.84
0.2821	14.0	210	0.4550	0.84
0.2713	15.0	225	0.4461	0.8343
0.2402	16.0	240	0.3447	0.8971
0.1729	17.0	255	0.3471	0.8914
0.1773	18.0	270	0.3235	0.8914
0.155	19.0	285	0.3583	0.8857
0.1867	20.0	300	0.3917	0.88
0.159	21.0	315	0.3539	0.8743
0.1498	22.0	330	0.3438	0.8629
0.1585	23.0	345	0.3092	0.8914
0.1655	24.0	360	0.2937	0.9029
0.1191	25.0	375	0.3470	0.8971
0.1756	26.0	390	0.3520	0.8857
0.1467	27.0	405	0.3552	0.88
0.1788	28.0	420	0.3994	0.8914
0.108	29.0	435	0.3685	0.9143
0.1359	30.0	450	0.3831	0.9086
0.0982	31.0	465	0.4162	0.8971
0.1154	32.0	480	0.3634	0.9086
0.0959	33.0	495	0.3828	0.8914
0.1136	34.0	510	0.4778	0.88
0.1184	35.0	525	0.3773	0.8857
0.0858	36.0	540	0.3628	0.8971
0.0958	37.0	555	0.4379	0.8857
0.0891	38.0	570	0.3993	0.8857
0.1132	39.0	585	0.3931	0.8857
0.093	40.0	600	0.3840	0.9314
0.094	41.0	615	0.3870	0.9029
0.0938	42.0	630	0.4061	0.8914
0.1024	43.0	645	0.5408	0.8571
0.0835	44.0	660	0.3230	0.9086
0.0795	45.0	675	0.3599	0.9086
0.053	46.0	690	0.4675	0.8686
0.0673	47.0	705	0.4926	0.8857
0.0888	48.0	720	0.4276	0.8971
0.0539	49.0	735	0.5204	0.8686
0.0662	50.0	750	0.4685	0.9143
0.0738	51.0	765	0.3824	0.9143
0.0459	52.0	780	0.4200	0.9143
0.0651	53.0	795	0.4323	0.9086
0.0744	54.0	810	0.4256	0.9086
0.0499	55.0	825	0.4435	0.9143
0.076	56.0	840	0.4884	0.9086
0.0689	57.0	855	0.4900	0.9029
0.0899	58.0	870	0.5848	0.8971
0.054	59.0	885	0.5400	0.9143
0.0524	60.0	900	0.5701	0.8914
0.0532	61.0	915	0.5269	0.9029
0.0596	62.0	930	0.5223	0.88
0.0814	63.0	945	0.5199	0.8914
0.0899	64.0	960	0.4566	0.9143
0.0649	65.0	975	0.4577	0.92
0.0508	66.0	990	0.4655	0.92
0.0529	67.0	1005	0.4868	0.9029
0.0707	68.0	1020	0.4883	0.9029
0.0603	69.0	1035	0.5059	0.9029
0.0776	70.0	1050	0.5878	0.8857
0.0747	71.0	1065	0.4694	0.92
0.0681	72.0	1080	0.5228	0.8857
0.0357	73.0	1095	0.5152	0.8857
0.049	74.0	1110	0.5197	0.9086
0.0335	75.0	1125	0.5028	0.9143
0.0523	76.0	1140	0.6165	0.8857
0.0258	77.0	1155	0.5491	0.8971
0.0398	78.0	1170	0.5218	0.8971
0.0408	79.0	1185	0.6784	0.8743
0.0606	80.0	1200	0.4997	0.9086
0.062	81.0	1215	0.5853	0.8857
0.0485	82.0	1230	0.5230	0.9086
0.0337	83.0	1245	0.5579	0.8971
0.0474	84.0	1260	0.5233	0.8971
0.0542	85.0	1275	0.5082	0.9143
0.0512	86.0	1290	0.5149	0.9029
0.0497	87.0	1305	0.5065	0.8971
0.0317	88.0	1320	0.4929	0.9143
0.0507	89.0	1335	0.5252	0.9143
0.0229	90.0	1350	0.5545	0.8914
0.0248	91.0	1365	0.5472	0.8914
0.0499	92.0	1380	0.5360	0.8971
0.0241	93.0	1395	0.5291	0.8971
0.0458	94.0	1410	0.5110	0.9029
0.0304	95.0	1425	0.5070	0.9029
0.1043	96.0	1440	0.5084	0.9086
0.0454	97.0	1455	0.5128	0.9086
0.0474	98.0	1470	0.5270	0.8971
0.0325	99.0	1485	0.5307	0.8971
0.0273	100.0	1500	0.5334	0.8971

Framework versions

• Transformers 4.51.3
• Pytorch 2.6.0+cu124
• Datasets 2.19.0
• Tokenizers 0.21.1