MTAL: A Novel Chinese Herbal Medicine Classification Approach with Mutual Triplet Attention Learning

Table 4

Comparisons of different models with or without mutual learning based on ResNet18 and ResNet50.

Model	Parameters	FLOPs	Accuracy
SigRes18	11.69 M	1.82 G	74.05
SigRes50	25.56 M	4.12 G	76.90
MulRes18(50)	11.69 M	1.82 G	78.80
MulRes50(18)	25.56 M	4.12 G	80.70