Skip to main content

Table 2 Network training hyperparameter setting

From: Improving multi-scale detection layers in the deep learning network for wheat spike detection based on interpretive analysis

Input size

Batch size

Epoch

Learning rate

Momentum

150 × 150

8

100

0.01

0.9

300 × 300

8

100

0.01

0.9

450 × 450

8

100

0.01

0.9

600 × 600

8

100

0.01

0.9