swin-tiny-patch4-window7-224-finetuned-st-wsdmhar-xyz-auc

This model is a fine-tuned version of microsoft/swin-tiny-patch4-window7-224 on the imagefolder dataset. It achieves the following results on the evaluation set:

• Loss: 0.1060
• Accuracy: 0.9766

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: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 100

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
1.586	1.0	53	1.4366	0.4559
0.9024	2.0	106	0.7526	0.7018
0.6157	3.0	159	0.5375	0.7944
0.5165	4.0	212	0.4298	0.8306
0.4315	5.0	265	0.3646	0.8609
0.3687	6.0	318	0.3054	0.8877
0.3352	7.0	371	0.2822	0.9005
0.3186	8.0	424	0.2764	0.9043
0.3155	9.0	477	0.2409	0.9215
0.2824	10.0	530	0.2459	0.9236
0.2575	11.0	583	0.2346	0.9129
0.2384	12.0	636	0.2445	0.9012
0.2117	13.0	689	0.1838	0.9342
0.2172	14.0	742	0.1789	0.9384
0.1918	15.0	795	0.1615	0.9480
0.1909	16.0	848	0.1516	0.9473
0.1911	17.0	901	0.1513	0.9494
0.2124	18.0	954	0.1524	0.9494
0.1631	19.0	1007	0.1729	0.9339
0.1766	20.0	1060	0.1329	0.9539
0.168	21.0	1113	0.1235	0.9590
0.1227	22.0	1166	0.1390	0.9483
0.1705	23.0	1219	0.1290	0.9566
0.1296	24.0	1272	0.1119	0.9621
0.1201	25.0	1325	0.1452	0.9497
0.1233	26.0	1378	0.1440	0.9487
0.1412	27.0	1431	0.1206	0.9573
0.1031	28.0	1484	0.1235	0.9580
0.1131	29.0	1537	0.1377	0.9501
0.1157	30.0	1590	0.1308	0.9580
0.0925	31.0	1643	0.1172	0.9601
0.0864	32.0	1696	0.1135	0.9621
0.0748	33.0	1749	0.0987	0.9656
0.1004	34.0	1802	0.0924	0.9728
0.0858	35.0	1855	0.1058	0.9659
0.0976	36.0	1908	0.1180	0.9587
0.0797	37.0	1961	0.1035	0.9676
0.0884	38.0	2014	0.0909	0.9707
0.0841	39.0	2067	0.0979	0.9707
0.0633	40.0	2120	0.0943	0.9697
0.0601	41.0	2173	0.1017	0.9687
0.0693	42.0	2226	0.1160	0.9652
0.0715	43.0	2279	0.0980	0.9704
0.0807	44.0	2332	0.1030	0.9711
0.0614	45.0	2385	0.0999	0.9707
0.0639	46.0	2438	0.1265	0.9632
0.0623	47.0	2491	0.1195	0.9614
0.0444	48.0	2544	0.1338	0.9659
0.0551	49.0	2597	0.1042	0.9728
0.0588	50.0	2650	0.0987	0.9731
0.0421	51.0	2703	0.1306	0.9607
0.0446	52.0	2756	0.1035	0.9718
0.0489	53.0	2809	0.1084	0.9714
0.0529	54.0	2862	0.1225	0.9663
0.0403	55.0	2915	0.1053	0.9711
0.0455	56.0	2968	0.1436	0.9645
0.0436	57.0	3021	0.1052	0.9714
0.0416	58.0	3074	0.1132	0.9666
0.0378	59.0	3127	0.1055	0.9721
0.0545	60.0	3180	0.1166	0.9704
0.0315	61.0	3233	0.1073	0.9711
0.0433	62.0	3286	0.1012	0.9735
0.0577	63.0	3339	0.1117	0.9714
0.0369	64.0	3392	0.1150	0.9697
0.0459	65.0	3445	0.1054	0.9731
0.0458	66.0	3498	0.1045	0.9745
0.0374	67.0	3551	0.1105	0.9725
0.0318	68.0	3604	0.1138	0.9718
0.0337	69.0	3657	0.1053	0.9728
0.0337	70.0	3710	0.1011	0.9738
0.0329	71.0	3763	0.1067	0.9738
0.0313	72.0	3816	0.1003	0.9756
0.0446	73.0	3869	0.1125	0.9714
0.047	74.0	3922	0.1040	0.9707
0.0256	75.0	3975	0.1165	0.9700
0.0535	76.0	4028	0.1129	0.9697
0.029	77.0	4081	0.1040	0.9752
0.044	78.0	4134	0.1116	0.9718
0.0405	79.0	4187	0.1130	0.9725
0.0417	80.0	4240	0.1094	0.9735
0.0257	81.0	4293	0.1143	0.9697
0.0293	82.0	4346	0.1111	0.9735
0.0234	83.0	4399	0.1253	0.9704
0.0295	84.0	4452	0.1133	0.9749
0.0261	85.0	4505	0.1048	0.9738
0.0215	86.0	4558	0.1072	0.9728
0.0304	87.0	4611	0.1061	0.9731
0.02	88.0	4664	0.1072	0.9742
0.0353	89.0	4717	0.1096	0.9738
0.0317	90.0	4770	0.1097	0.9745
0.0441	91.0	4823	0.1080	0.9745
0.0262	92.0	4876	0.1051	0.9752
0.0312	93.0	4929	0.1089	0.9738
0.025	94.0	4982	0.1094	0.9738
0.0243	95.0	5035	0.1106	0.9745
0.0245	96.0	5088	0.1076	0.9752
0.0233	97.0	5141	0.1068	0.9762
0.0279	98.0	5194	0.1063	0.9759
0.0285	99.0	5247	0.1057	0.9766
0.022	100.0	5300	0.1060	0.9766

Framework versions

• Transformers 4.44.2
• Pytorch 2.5.0.dev20240829+cu118
• Datasets 2.19.2
• Tokenizers 0.19.1