Table 5 NHANet model performance metrics at a learning rate of 0.0001.

From: An incremental adversarial training method enables timeliness and rapid new knowledge acquisition

Batch-size

8

16

32

64

128

Accuracy

0.9773

0.9693

0.9053

0.8160

0.7520

Precision

0.9772

0.9694

0.9032

0.8111

0.7494

Recall

0.9772

0.9703

0.9035

0.8134

0.7484

F1-score

0.9771

0.9694

0.9026

0.8106

0.7444

Execution-time

487.1841s

256.0790 s

143.1442 s

90.4628 s

64.0669 s

Back to article page