Skip to main content
. 2020 Jul 25;5(3):509–518. doi: 10.1007/s41403-020-00157-z

Table 7.

mAP and FPS of trained models on Moxa3K dataset

Model mAP@50 FPS
YOLOv3 414x414 63.99 21.2
YOLOv3 608x608 66.84 10.9
YOLOv3 832x832 61.73 6.9
YOLOv3Tiny 414x414 56.27 138
YOLOv3Tiny 608x608 55.08 72
YOLOv3Tiny 832x832 56.57 46.5
SSD 300 MobliNetv2 46.52 67.1
F-RCNN 300 Inceptionv2 60.5 14.8