Skip to main content
. 2024 Apr 18;15:1373590. doi: 10.3389/fpls.2024.1373590

Table 7.

Performance summary of YOLOv8s with additional 3 Conv blocks and alternative activation functions on the validation dataset.

Model Precision
val | test
Recall
val | test
mAP50
val | test
mAP50-95
val | test
YOLOv8s+Conv3+SiLU 0.899 | 0.931 0.846 | 0.829 0.904 | 0.906 0.674 | 0.694
YOLOv8s+Conv3+ReLU 0.905 | 0.906 0.815 | 0.829 0.878 | 0.875 0.626 | 0.668
YOLOv8s+Conv3+LeakyReLU 0.915 | 0.942 0.834 | 0.840 0.892 | 0.900 0.642 | 0.677
YOLOv8s+Conv3+Tanh 0.822 | 0.906 0.623 | 0.702 0.726 | 0.771 0.434 | 0.502
YOLOv8s+Conv3+Hard Swish 0.919 | 0.932 0.851 | 0.826 0.920 | 0.911 0.677 | 0.701

Bold values signify the value of the best performing activation function configuration under each metric.