Edit model card

swin-tiny-patch4-window7-224-finetuned-st-wsdmhar-160

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: 0.1319
  • Accuracy: 0.9735

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: 160

Training results

Training Loss Epoch Step Validation Loss Accuracy
1.6426 1.0 53 1.5279 0.4642
0.9084 2.0 106 0.7447 0.7028
0.6514 3.0 159 0.5340 0.7934
0.56 4.0 212 0.5470 0.7621
0.4578 5.0 265 0.3704 0.8547
0.4811 6.0 318 0.3317 0.8705
0.4211 7.0 371 0.3045 0.8764
0.3828 8.0 424 0.3240 0.8908
0.3505 9.0 477 0.2840 0.8836
0.3652 10.0 530 0.2430 0.9084
0.3385 11.0 583 0.2359 0.9153
0.3071 12.0 636 0.2899 0.8946
0.3319 13.0 689 0.2588 0.9108
0.2657 14.0 742 0.1920 0.9270
0.2423 15.0 795 0.1830 0.9370
0.2625 16.0 848 0.2116 0.9273
0.2394 17.0 901 0.1917 0.9353
0.2245 18.0 954 0.1894 0.9280
0.2123 19.0 1007 0.1768 0.9346
0.2158 20.0 1060 0.1902 0.9311
0.1943 21.0 1113 0.1704 0.9390
0.1712 22.0 1166 0.1442 0.9466
0.2073 23.0 1219 0.1279 0.9552
0.1676 24.0 1272 0.1548 0.9459
0.1775 25.0 1325 0.1371 0.9497
0.1644 26.0 1378 0.1247 0.9563
0.1652 27.0 1431 0.1547 0.9408
0.1383 28.0 1484 0.1301 0.9545
0.1268 29.0 1537 0.1484 0.9504
0.1252 30.0 1590 0.1385 0.9549
0.1288 31.0 1643 0.1368 0.9521
0.1353 32.0 1696 0.1184 0.9597
0.1585 33.0 1749 0.1218 0.9570
0.1445 34.0 1802 0.1173 0.9597
0.1381 35.0 1855 0.1160 0.9614
0.1398 36.0 1908 0.1292 0.9542
0.1138 37.0 1961 0.1135 0.9597
0.147 38.0 2014 0.0958 0.9690
0.0927 39.0 2067 0.1008 0.9663
0.1031 40.0 2120 0.1105 0.9659
0.1197 41.0 2173 0.1010 0.9656
0.1325 42.0 2226 0.1178 0.9621
0.0862 43.0 2279 0.1042 0.9659
0.1037 44.0 2332 0.1016 0.9680
0.0885 45.0 2385 0.1063 0.9649
0.1217 46.0 2438 0.1117 0.9673
0.0947 47.0 2491 0.1048 0.9673
0.0831 48.0 2544 0.1061 0.9666
0.1082 49.0 2597 0.0946 0.9680
0.0856 50.0 2650 0.1139 0.9694
0.0832 51.0 2703 0.1152 0.9618
0.0823 52.0 2756 0.0970 0.9721
0.0773 53.0 2809 0.1049 0.9683
0.0794 54.0 2862 0.1048 0.9731
0.0813 55.0 2915 0.1089 0.9669
0.079 56.0 2968 0.0982 0.9704
0.095 57.0 3021 0.1242 0.9680
0.0775 58.0 3074 0.1262 0.9676
0.0795 59.0 3127 0.1276 0.9649
0.0619 60.0 3180 0.0937 0.9704
0.0688 61.0 3233 0.1149 0.9707
0.0932 62.0 3286 0.1019 0.9700
0.0675 63.0 3339 0.1239 0.9687
0.0715 64.0 3392 0.1143 0.9669
0.0858 65.0 3445 0.1053 0.9680
0.0646 66.0 3498 0.1150 0.9694
0.0736 67.0 3551 0.1119 0.9700
0.0665 68.0 3604 0.1031 0.9721
0.0509 69.0 3657 0.1069 0.9731
0.0642 70.0 3710 0.1171 0.9704
0.0588 71.0 3763 0.1235 0.9718
0.0837 72.0 3816 0.1121 0.9700
0.0534 73.0 3869 0.1162 0.9704
0.0612 74.0 3922 0.1116 0.9697
0.0621 75.0 3975 0.1220 0.9700
0.063 76.0 4028 0.1084 0.9714
0.0604 77.0 4081 0.1180 0.9694
0.0511 78.0 4134 0.1325 0.9687
0.05 79.0 4187 0.1179 0.9680
0.072 80.0 4240 0.1516 0.9597
0.0746 81.0 4293 0.1159 0.9714
0.0544 82.0 4346 0.1201 0.9707
0.0527 83.0 4399 0.1232 0.9725
0.044 84.0 4452 0.1450 0.9700
0.0462 85.0 4505 0.1229 0.9690
0.0445 86.0 4558 0.1404 0.9669
0.0524 87.0 4611 0.1153 0.9711
0.0638 88.0 4664 0.1207 0.9707
0.0435 89.0 4717 0.1289 0.9718
0.0567 90.0 4770 0.1167 0.9700
0.0553 91.0 4823 0.1100 0.9742
0.0566 92.0 4876 0.1319 0.9721
0.0462 93.0 4929 0.1275 0.9707
0.0539 94.0 4982 0.1263 0.9711
0.0561 95.0 5035 0.1333 0.9725
0.0362 96.0 5088 0.1241 0.9704
0.0435 97.0 5141 0.1199 0.9714
0.0637 98.0 5194 0.1290 0.9707
0.0466 99.0 5247 0.1200 0.9666
0.0471 100.0 5300 0.1556 0.9656
0.0407 101.0 5353 0.1334 0.9707
0.0375 102.0 5406 0.1307 0.9707
0.0375 103.0 5459 0.1392 0.9687
0.0354 104.0 5512 0.1237 0.9714
0.0523 105.0 5565 0.1298 0.9711
0.0307 106.0 5618 0.1283 0.9687
0.0427 107.0 5671 0.1300 0.9683
0.0327 108.0 5724 0.1292 0.9711
0.0411 109.0 5777 0.1377 0.9683
0.0422 110.0 5830 0.1260 0.9697
0.044 111.0 5883 0.1183 0.9731
0.0332 112.0 5936 0.1347 0.9735
0.0302 113.0 5989 0.1251 0.9731
0.0273 114.0 6042 0.1100 0.9728
0.0442 115.0 6095 0.1368 0.9728
0.0337 116.0 6148 0.1308 0.9697
0.0395 117.0 6201 0.1198 0.9738
0.0398 118.0 6254 0.1344 0.9697
0.0362 119.0 6307 0.1243 0.9752
0.035 120.0 6360 0.1363 0.9735
0.0389 121.0 6413 0.1271 0.9756
0.0305 122.0 6466 0.1277 0.9759
0.0366 123.0 6519 0.1276 0.9704
0.0329 124.0 6572 0.1192 0.9780
0.0304 125.0 6625 0.1325 0.9728
0.0289 126.0 6678 0.1334 0.9728
0.0362 127.0 6731 0.1272 0.9707
0.0326 128.0 6784 0.1250 0.9735
0.0357 129.0 6837 0.1255 0.9749
0.0264 130.0 6890 0.1326 0.9769
0.0324 131.0 6943 0.1359 0.9752
0.0258 132.0 6996 0.1229 0.9766
0.033 133.0 7049 0.1184 0.9759
0.0259 134.0 7102 0.1416 0.9718
0.0362 135.0 7155 0.1310 0.9745
0.0263 136.0 7208 0.1434 0.9728
0.0406 137.0 7261 0.1271 0.9745
0.027 138.0 7314 0.1395 0.9728
0.0417 139.0 7367 0.1307 0.9735
0.0321 140.0 7420 0.1276 0.9742
0.0451 141.0 7473 0.1338 0.9759
0.029 142.0 7526 0.1337 0.9749
0.0337 143.0 7579 0.1315 0.9745
0.0212 144.0 7632 0.1331 0.9759
0.0301 145.0 7685 0.1291 0.9759
0.0306 146.0 7738 0.1276 0.9749
0.0283 147.0 7791 0.1275 0.9731
0.0291 148.0 7844 0.1293 0.9752
0.0265 149.0 7897 0.1381 0.9749
0.0326 150.0 7950 0.1308 0.9742
0.0301 151.0 8003 0.1279 0.9731
0.021 152.0 8056 0.1312 0.9735
0.0186 153.0 8109 0.1364 0.9735
0.0322 154.0 8162 0.1367 0.9725
0.0229 155.0 8215 0.1347 0.9745
0.0249 156.0 8268 0.1360 0.9728
0.0312 157.0 8321 0.1325 0.9731
0.0295 158.0 8374 0.1315 0.9735
0.0234 159.0 8427 0.1308 0.9738
0.0239 160.0 8480 0.1319 0.9735

Framework versions

  • Transformers 4.43.2
  • Pytorch 2.3.1+cu118
  • Datasets 2.20.0
  • Tokenizers 0.19.1
Downloads last month
3
Safetensors
Model size
27.6M params
Tensor type
I64
·
F32
·
Inference Examples
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social visibility and check back later, or deploy to Inference Endpoints (dedicated) instead.

Model tree for ayubkfupm/swin-tiny-patch4-window7-224-finetuned-st-wsdmhar-160

Finetuned
(440)
this model

Evaluation results