Table 5 Intermediate results of Model+ in Expt.#4

From: Investigating the key principles in two-step heterogeneous transfer learning for early laryngeal cancer identification

Model

Epoch

Val. Acc.1

Avg. Time 2

Model

Epoch

Val. Acc.

Avg. Time

DenseNet121+

26

0.791

136s

ResNet50+

38

0.743

106s

DenseNet169+

26

0.777

156s

VGG19+

32

0.738

150s

InceptionV3+

20

0.822

142s

ViT-B/16+

15

0.759

237s

MobileNet V2+

17

0.808

120s

EfficientNetV2-S+

13

0.84

251s

ResNet18+

19

0.773

121s

SwinV2-T+

16

0.825

543s

MaxVit-T+

13

0.836

548s

    
  1. 1Validation accuracy.2 Average time per epoch, and is rounded up
Back to article page