swin-tiny-patch4-window7-224-finetuned-st-wsdmhar-stacked_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: 0.1284
• Accuracy: 0.9656

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
1.452	1.0	53	1.2923	0.5028
0.7902	2.0	106	0.6596	0.7262
0.6011	3.0	159	0.5092	0.7889
0.5202	4.0	212	0.4173	0.8361
0.4687	5.0	265	0.3498	0.8574
0.422	6.0	318	0.3566	0.8633
0.3678	7.0	371	0.3017	0.8877
0.3414	8.0	424	0.2601	0.9050
0.3249	9.0	477	0.2777	0.8884
0.3136	10.0	530	0.2625	0.9063
0.3163	11.0	583	0.2504	0.9063
0.3144	12.0	636	0.2146	0.9215
0.3062	13.0	689	0.2225	0.9198
0.2574	14.0	742	0.1928	0.9304
0.2487	15.0	795	0.1974	0.9236
0.2022	16.0	848	0.2147	0.9215
0.198	17.0	901	0.1708	0.9394
0.2174	18.0	954	0.2144	0.9156
0.225	19.0	1007	0.1866	0.9360
0.1886	20.0	1060	0.1512	0.9439
0.1797	21.0	1113	0.1520	0.9473
0.1779	22.0	1166	0.1944	0.9294
0.2006	23.0	1219	0.1722	0.9404
0.1647	24.0	1272	0.1401	0.9487
0.1766	25.0	1325	0.1587	0.9449
0.1347	26.0	1378	0.1525	0.9504
0.1533	27.0	1431	0.1336	0.9528
0.1322	28.0	1484	0.2079	0.9329
0.1291	29.0	1537	0.1421	0.9518
0.1397	30.0	1590	0.1457	0.9497
0.1189	31.0	1643	0.1530	0.9521
0.1404	32.0	1696	0.1818	0.9332
0.1431	33.0	1749	0.1486	0.9487
0.1214	34.0	1802	0.1555	0.9525
0.1195	35.0	1855	0.1852	0.9439
0.1161	36.0	1908	0.1670	0.9439
0.1052	37.0	1961	0.1551	0.9504
0.1004	38.0	2014	0.1535	0.9511
0.113	39.0	2067	0.1308	0.9514
0.114	40.0	2120	0.1752	0.9463
0.0807	41.0	2173	0.1467	0.9528
0.1044	42.0	2226	0.1289	0.9604
0.1118	43.0	2279	0.1602	0.9518
0.1305	44.0	2332	0.1699	0.9452
0.083	45.0	2385	0.1376	0.9563
0.1153	46.0	2438	0.1272	0.9594
0.0875	47.0	2491	0.1358	0.9559
0.0772	48.0	2544	0.1662	0.9501
0.084	49.0	2597	0.1456	0.9580
0.082	50.0	2650	0.1593	0.9483
0.0919	51.0	2703	0.1638	0.9483
0.0999	52.0	2756	0.1420	0.9532
0.0718	53.0	2809	0.1447	0.9549
0.0757	54.0	2862	0.1791	0.9490
0.0632	55.0	2915	0.1364	0.9604
0.0922	56.0	2968	0.1544	0.9525
0.0805	57.0	3021	0.1493	0.9552
0.0702	58.0	3074	0.1307	0.9570
0.0554	59.0	3127	0.1502	0.9532
0.0699	60.0	3180	0.1340	0.9590
0.0759	61.0	3233	0.1353	0.9576
0.0604	62.0	3286	0.1441	0.9570
0.0642	63.0	3339	0.1312	0.9601
0.0577	64.0	3392	0.1399	0.9597
0.0506	65.0	3445	0.1347	0.9594
0.0781	66.0	3498	0.1403	0.9601
0.0664	67.0	3551	0.1379	0.9587
0.0775	68.0	3604	0.1389	0.9573
0.0578	69.0	3657	0.1360	0.9570
0.0782	70.0	3710	0.1317	0.9580
0.0474	71.0	3763	0.1446	0.9594
0.0357	72.0	3816	0.1359	0.9618
0.0472	73.0	3869	0.1429	0.9590
0.071	74.0	3922	0.1333	0.9604
0.0663	75.0	3975	0.1327	0.9597
0.0536	76.0	4028	0.1396	0.9587
0.0549	77.0	4081	0.1392	0.9597
0.0621	78.0	4134	0.1408	0.9645
0.0531	79.0	4187	0.1406	0.9607
0.0464	80.0	4240	0.1463	0.9594
0.0526	81.0	4293	0.1355	0.9638
0.0277	82.0	4346	0.1464	0.9635
0.0558	83.0	4399	0.1487	0.9604
0.0466	84.0	4452	0.1319	0.9642
0.0463	85.0	4505	0.1443	0.9621
0.0397	86.0	4558	0.1494	0.9611
0.0489	87.0	4611	0.1428	0.9642
0.0354	88.0	4664	0.1387	0.9645
0.0457	89.0	4717	0.1362	0.9638
0.0522	90.0	4770	0.1332	0.9656
0.0481	91.0	4823	0.1352	0.9642
0.0472	92.0	4876	0.1375	0.9673
0.0362	93.0	4929	0.1354	0.9656
0.0432	94.0	4982	0.1306	0.9632
0.037	95.0	5035	0.1283	0.9663
0.0525	96.0	5088	0.1273	0.9666
0.0349	97.0	5141	0.1279	0.9659
0.0411	98.0	5194	0.1279	0.9659
0.044	99.0	5247	0.1283	0.9659
0.0289	100.0	5300	0.1284	0.9656

Framework versions

• Transformers 4.44.0
• Pytorch 1.12.1+cu113
• Datasets 2.21.0
• Tokenizers 0.19.1