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

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: 0.1144
• Accuracy: 0.9711

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
1.4641	1.0	53	1.2510	0.5768
0.7713	2.0	106	0.6282	0.7435
0.5822	3.0	159	0.4503	0.8251
0.5086	4.0	212	0.4028	0.8306
0.4499	5.0	265	0.3399	0.8709
0.4171	6.0	318	0.3024	0.8898
0.3597	7.0	371	0.2584	0.9029
0.3201	8.0	424	0.2467	0.9143
0.2974	9.0	477	0.2784	0.8981
0.3323	10.0	530	0.2414	0.9101
0.2717	11.0	583	0.2051	0.9349
0.2647	12.0	636	0.1944	0.9294
0.296	13.0	689	0.1871	0.9329
0.2434	14.0	742	0.1701	0.9411
0.2293	15.0	795	0.1685	0.9435
0.2196	16.0	848	0.1486	0.9446
0.2249	17.0	901	0.1438	0.9456
0.2142	18.0	954	0.1529	0.9449
0.2024	19.0	1007	0.1361	0.9532
0.2177	20.0	1060	0.1906	0.9315
0.1812	21.0	1113	0.1333	0.9532
0.1771	22.0	1166	0.1424	0.9528
0.1718	23.0	1219	0.1370	0.9545
0.1333	24.0	1272	0.1550	0.9466
0.1592	25.0	1325	0.1344	0.9535
0.1279	26.0	1378	0.1201	0.9590
0.1547	27.0	1431	0.1225	0.9597
0.1443	28.0	1484	0.1526	0.9480
0.1306	29.0	1537	0.1188	0.9597
0.129	30.0	1590	0.1204	0.9614
0.1354	31.0	1643	0.1351	0.9570
0.135	32.0	1696	0.1010	0.9676
0.137	33.0	1749	0.1381	0.9566
0.101	34.0	1802	0.1119	0.9642
0.1118	35.0	1855	0.1056	0.9656
0.1206	36.0	1908	0.0975	0.9663
0.1028	37.0	1961	0.1265	0.9635
0.1058	38.0	2014	0.0958	0.9680
0.1013	39.0	2067	0.1060	0.9642
0.0765	40.0	2120	0.1024	0.9659
0.0997	41.0	2173	0.1116	0.9642
0.0933	42.0	2226	0.1082	0.9666
0.0937	43.0	2279	0.1095	0.9680
0.0831	44.0	2332	0.1034	0.9683
0.0849	45.0	2385	0.0995	0.9690
0.0739	46.0	2438	0.1042	0.9666
0.0944	47.0	2491	0.1068	0.9697
0.0821	48.0	2544	0.1087	0.9690
0.0835	49.0	2597	0.0975	0.9721
0.0666	50.0	2650	0.1189	0.9638
0.0614	51.0	2703	0.1421	0.9611
0.0738	52.0	2756	0.1253	0.9638
0.0949	53.0	2809	0.1274	0.9663
0.068	54.0	2862	0.1051	0.9669
0.0626	55.0	2915	0.1102	0.9673
0.0647	56.0	2968	0.1096	0.9673
0.0803	57.0	3021	0.1049	0.9683
0.0744	58.0	3074	0.1039	0.9697
0.0769	59.0	3127	0.1060	0.9690
0.0763	60.0	3180	0.1077	0.9680
0.0591	61.0	3233	0.1165	0.9680
0.0649	62.0	3286	0.1109	0.9694
0.0557	63.0	3339	0.1162	0.9680
0.0644	64.0	3392	0.1039	0.9718
0.0558	65.0	3445	0.1182	0.9687
0.0633	66.0	3498	0.1151	0.9680
0.0586	67.0	3551	0.1147	0.9694
0.0651	68.0	3604	0.1124	0.9711
0.0693	69.0	3657	0.1104	0.9687
0.0584	70.0	3710	0.1177	0.9697
0.0471	71.0	3763	0.1160	0.9690
0.0614	72.0	3816	0.1220	0.9680
0.0583	73.0	3869	0.1236	0.9656
0.0495	74.0	3922	0.1076	0.9718
0.0574	75.0	3975	0.1163	0.9673
0.0399	76.0	4028	0.1126	0.9683
0.0357	77.0	4081	0.1064	0.9728
0.0441	78.0	4134	0.1139	0.9694
0.0504	79.0	4187	0.1083	0.9707
0.0546	80.0	4240	0.1167	0.9676
0.0528	81.0	4293	0.1143	0.9697
0.0385	82.0	4346	0.1226	0.9676
0.0511	83.0	4399	0.1199	0.9694
0.0533	84.0	4452	0.1279	0.9673
0.043	85.0	4505	0.1161	0.9714
0.0231	86.0	4558	0.1166	0.9728
0.0426	87.0	4611	0.1239	0.9690
0.0565	88.0	4664	0.1189	0.9687
0.0378	89.0	4717	0.1186	0.9697
0.0406	90.0	4770	0.1209	0.9718
0.0306	91.0	4823	0.1189	0.9721
0.0354	92.0	4876	0.1244	0.9687
0.0293	93.0	4929	0.1235	0.9697
0.0381	94.0	4982	0.1186	0.9711
0.0372	95.0	5035	0.1172	0.9714
0.0469	96.0	5088	0.1180	0.9711
0.0535	97.0	5141	0.1152	0.9718
0.0496	98.0	5194	0.1157	0.9714
0.034	99.0	5247	0.1145	0.9714
0.0348	100.0	5300	0.1144	0.9711

Framework versions

• Transformers 4.43.2
• Pytorch 2.3.1+cu118
• Datasets 2.20.0
• Tokenizers 0.19.1