YOLOv4推理時間是v3的2.5倍

显示全部楼层 · 发表于 2020-10-9 16:34:33

I believe the reason for yolov4 being so much more slower is the "mish" activation layer. since it uses exponential calculations, it could be behind this bottleneck.

I am trying to find a reasonable solution to this issue as well, if you have any idea please share with me mate.

cheers

显示全部楼层 · 发表于 2020-10-20 20:31:08

Ohh I see, so the solution would be to remove those pooling layers from base network and then retrain with imagenet?

显示全部楼层 · 发表于 2020-11-13 15:54:36

Did you manage to find a solution to this issue? I asked a way to remove spp in darknet repository, but haven't got an answer.

I wonder why spp structure causes a bottleneck. Normally maxpool layers work fine, but when in spp they are taking too much time.